This is not a DeFi story. It happened entirely within traditional private credit, with all of its existing disclosure infrastructure. But it is exactly the right place to start, because it shows the gap between what blockchains promise and what we currently have: on a public ledger, the deterioration of that position could have been visible in real time, and today it is not. That gap between promised transparency and actual market structure is the thread that runs through everything that follows: DeFi, tokenization, stablecoins, and why distribution and yield still do not line up with the story we tell about “on-chain finance.”

Subscribe now

I. What DeFi Actually Does

For those who still do not truly understand the topic: DeFi is a set of financial protocols running on public blockchains that allow anyone with a wallet to lend, borrow, and trade without a bank or broker in the middle. No account minimums. No compliance officer deciding who qualifies. Automated settlement on‑chain.

It has three functions worth taking seriously.

The first is democratizing access to capital markets. Anyone with a wallet can access and borrow via Aave, or create entirely new lending markets on Morpho without permission from anyone. Crypto‑native trading firms that get rejected by banks, not because of credit quality but because banks do not want sector exposure, can access global liquidity pools through platforms like Maple at rates between 6–12%, with collateral ratios of 120–170%. For borrowers that traditional credit will not touch, it is often the only option. Moreover, protocols like Fluid offer self‑repaying yield positions, USD.ai brings new collateral types on‑chain, and Gauntlet and Steakhouse run multi‑strategy vaults that automatically optimize across protocols. None of this infrastructure exists in traditional capital markets at anything close to this level of accessibility.

The second is creating more liquid capital markets. Thousands of participants lend from the same pool and redeem when they want. The yield‑bearing positions they hold, essentially IOUs, can be traded on secondary markets and used as collateral in other protocols. Capital becomes fungible in a way it simply is not in traditional capital markets, where many positions are illiquid, exits are slow, and each instrument is a bilateral negotiation.

The third is improving yield diversity. DeFi allows participants to explicitly choose their risk position within the same underlying asset, something that has no real parallel in traditional markets. Tranching lets you select fixed or variable yield. Looping, using your yield‑bearing position as collateral to borrow more, lets you increase leverage instantly. Ethena’s sUSDe basis trade packages the funding‑rate spread into a single instrument. The gold‑backed thUSD does the same for commodities. Ethena’s Strata product lets you select your yield tranche explicitly: senior, mezzanine, or junior, from the same underlying pool.

This is risk repackaging, not unlike what CDOs did in 2004. Pendle separates principal and yield into tradeable tokens. Stablecoins span a spectrum from T‑bill‑backed USDC to private‑credit‑backed syrupUSDC. Looping adds explicit leverage. Instead of three tranches, we now have tens of instruments representing different positions on the same risk curve.

The promise was that this time it would be different from 2008. Instead of having to be , manually excavating prospectuses to find what was inside a mortgage-backed security, any wallet holder could verify the composition of their position in real time. The yield repackaging and the transparency were supposed to arrive together.

The yield repackaging arrived. The transparency has not (yet).

II. What Is Being Tokenized?

Capital flowing into tokenized assets has settled into a clear hierarchy since 2025. US Treasury debt has added the most in absolute terms, given the yield and safety. Commodities have added billions of dollars, driven almost entirely by renewed institutional interest in gold. The fastest‑growing segments in percentage terms, however, are corporate credit and diversified credit like Figure’s HELOC product, Apollo’s ACRED, and Arca’s Stable Return fund.

One distinction matters enormously and gets glossed over in almost every conversation about tokenization: some of these assets are represented on‑chain, not originated there. Figure’s HELOC (a tokenized home‑equity line of credit) is only represented on‑chain, but you can still use it as collateral on Kamino. The underlying capital‑markets infrastructure (the underwriting, legal enforceability, and servicing) remains off‑chain and centralized. The blockchain is functioning as a ledger, not a lending system.

III. The DeFi and Stablecoin Distribution Problem

A DeFi lending protocol’s economics are a function of four variables:

AUM × collateralization rate × utilization × borrow fees + establishment fees on new loans.

The lever that most directly determines health is utilization. And utilization is falling or plateauing even as headline AUM grows.

As retail and crypto‑native demand for borrowing gets tapped out, the next frontier is TradFi institutions. However, there are issues with this.

Maple’s data makes this visible. Total AUM has grown significantly, driven by syrupUSDC expanding rapidly into retail, but institutional pools have been relatively flat as a share of the whole. syrupUSDC’s AUM reached around $3 billion in 2025 with lower utilization than smaller, more specialized pools, while other syrup products frequently operated near full deployment. In other words, AUM growth driven by retail deposits has not been matched by institutional borrower demand.

Miners no longer want to borrow at scale. Large funds and family offices no longer need to, because there may be better risk‑return opportunities elsewhere. We see a similar trend in Aave’s institutional borrowing platform, Aave Horizon, where large on‑chain positions and DEX trade data confirm that institutional borrowing demand is no longer growing as it once did.

For this reason, DeFi protocols are in a search for marginal demand from borrowers who are less tied to crypto cycles. Protocols are turning to neo‑finance companies, bank partnerships, and private‑credit originators to fill this gap.

But reaching those borrowers requires consumer distribution infrastructure, and that is precisely what DeFi has never built. Aave has wound down several consumer‑facing products and is currently trying again with Aave App, which it is clearly pushing on its homepage over its prosumer DeFi dashboard.

This is also the reason why crypto credit cards have been all the rage among many DeFi (and other) protocols, including Ether.fi, Coinbase, MetaMask, 1inch, and Plasma. However, this does not solve the distribution issue; it has become more like table stakes than meaningful innovation. Crypto payment cards are not a new idea and have seen only limited adoption. The biggest beneficiary is Visa and sponsor banks, not the new stablecoin neobanks.

While there are debit/credit card reward nerds who compare card rewards, they are few and far between. Moreover, these cards are not likely to outcompete modern neobanks like Revolut, even with slightly higher APYs. Revolut already shares around 4% APYs with its card users and can negotiate strong packages for its users: free subscriptions to things like the Financial Times, Perplexity Pro, and NordVPN, zero FX fees, robo‑advisors, and sophisticated business‑card software. It has processed over $10 billion in stablecoin volume, versus KAST’s roughly $5 billion annualized.

The point is that without becoming an all‑in‑one app, having a card will not meaningfully increase deposits, outside a limited set of reward hackers. This is why, while most stablecoin startups are founded in the US or Europe, players like KAST and Tether are all looking primarily to the Global South. Not only do they not have to compete with entrenched neobanks to the same degree (though Nubank is an exception, and Revolut has made inroads into Mexico, Colombia, Argentina, and South Africa), they also do not have to compete with relatively well‑functioning domestic payment schemes in Europe and the US, where it is not obvious that stablecoins win on speed and cost.

The Nubank Cinderella story is appealing. Its LTV/CAC is roughly 10x that of Revolut, driven largely by much lower customer acquisition costs. It holds around 58% of LATAM neobank customers but likely generates 80–85% of the region’s neobank revenue. In places like LATAM, the high LTV/CAC ratio is driven by low CAC rather than unusually high LTV.

This poses two questions for stablecoin companies entering the region: (a) how much more in CAC will it cost to pry customers away from Nubank, and (b) if several stablecoin companies start going after LATAM clients, how much will CAC rise?

The beauty of DeFi for these entrenched neobanks (as opposed to stablecoin startups) is that curators like Gauntlet and Steakhouse make it simple for them to earn the same yields that stablecoin cards like KAST use. Down the line they could even spin up their own Morpho vault, pass on 100% of the yield to customers, and treat it as a loss leader. Taking a cut of such services (including yield) is a key revenue strategy for stablecoin startups or networks like Plasma, while incumbents can undercut those fees. It is not inconceivable that entire blockchains, like Tempo, could act as loss leaders to Stripe. Chris Dixon () has long argued that stablecoins and stablecoin transactions will become public goods, which is plausible.

Moreover, incumbent payment cards are objectively better products than stablecoin ones and likely have better margins, because they spend less on brand recognition and have banking licenses.

So if incumbents have better distribution and lower marketing costs, more mature ecosystems of services, and decent yield, how can stablecoin startups compete?

The realistic differentiators are:

• Easier off‑ramps for crypto traders and investors (but competition is catching up).

• Slightly higher yield via DeFi and token incentives (though token rewards are speculative, and most stablecoin card issuers do not even have a token yet).

• Higher cashback.

Delving deeper into KAST’s unit economics, KAST makes money in three main ways:

• Capturing a portion of KAST Earn yield (undisclosed cut).

• Taking 1–2% interchange fees (outside Europe), of which 20–50% go to its sponsor bank; by contrast, Revolut, N26, and Monzo have banking licenses and keep this spread.

• Charging miscellaneous fees like FX and declined‑transaction fees (both unusual for neobanks).

Modeled (below) across its three cards (entry, mid, and high end), with optimistic assumptions, the takeaways are:

• To maximize profit per cardholder, KAST would want users to deposit a lot and spend less; cashback is a loss leader when it exceeds net interchange.

• To maintain profitability, KAST is incentivized to withhold as much yield as possible, which pushes it down the risk curve if it still wants to beat neobank yields.

• Higher‑tier cards are more profitable; the entry‑level K‑Card is effectively a loss leader.

• LATAM and APAC are the most profitable regions for KAST.

Unlike incumbents, KAST also does not have business or credit cards, the highest‑LTV product categories. Around 85% of Nubank’s revenue comes from credit products.

To offer credit cards in Latin America, KAST would need banking‑level licenses per country, substantial capital, and enhanced risk systems, beyond its current fintech/sponsor‑bank model.

The once‑strong moats are dwindling as well. DeFi yields may get competed away as incumbent banks plug into curator vaults, and DeFi yield has always been downstream of one thing: on‑chain trading activity. Ethena’s basis trade depends on funding rates produced by leveraged traders. Borrowing demand at Aave and Maple correlates directly with how much trading is happening. When trading volume is high and volatile, the whole system runs. When it drops, every yield source downstream feels it.

Crypto on‑off‑ramping’s oligopoly is being attacked from both directions. From the crypto‑native side there is increased competition from MetaMask, Coinbase, Ether.fi, and Phantom. From the incumbent neobank side we see names like Robinhood and Revolut; competition is becoming saturated.

Higher fees work when you are the only game in town serving high‑net‑worth crypto investors and traders in wealthy geographies; once you go to the Global South and face economies of scale, this no longer works. Economies of scale matter for Visa transaction fees too. KAST likely has higher fees than its larger competitors due to lower volumes.

Crypto remains structurally important, but stablecoin players whose only value add is crypto on/off ramps will not succeed in the long term.

The only ways they can succeed are:

• Giving away generous token incentives while somehow avoiding a collapse in token price, which history suggests is unlikely to be sustainable.

• Offering genuinely differentiated APY from more exotic asset pools or income streams (sequencing fees, MEV, niche strategies), instead of simple USDC yield that is accessible to anyone. Here, vertical integration is more competitive than horizontal aggregation. Think Maple offering unique APY only to its cardholders, or a future Pump.Fun card allocating cashback directly to daily memecoin runners instead of paying out in depreciating dollars. Next‑generation stablecoin payment startups will likely offer APY sourced from venues like Hyperliquid, HIP‑3, and Polymarket trading strategies.

This naturally pushes them down the risk curve and toward outcomes like Stream or Goldfinch, where private‑credit risk and operational risk combine in ways that are hard to see until something breaks.

The bottom line is that yield is the decisive commodity, whether it is for DeFi protocols, DeFi curators, stablecoin payment cards, or bank savings accounts. DeFi protocols like Maple and Aave will live and die by the APY rates they can sustainably offer. Morpho seems more impervious as a pure platform where DeFi natives and Wall Street veterans can spin up custom pools, though it too will suffer if DeFi yields fail to materialize.

It is no surprise that financial institutions like Apollo are starting to wrap their tentacles around Morpho. It is plausible that Apollo will one day launch on‑chain private‑credit pools that are genuinely on‑chain rather than just represented there.

Ultimately there are two paths DeFi protocols and stablecoin platforms can choose: (1) trying to become a superapp, hoping that category‑invention benefits outweigh the distribution handicap (Aave, KAST, Phantom), or (2) becoming the middleware layer and hoping incumbents do not circumvent you (Morpho, Rain, Gauntlet).

IV. New Yield Categories

With AI and prediction markets funneling speculative capital away from leveraged crypto trading, it is clear DeFi needs to expand its sources of yield.

Private credit and other alternative markets (like GPU‑backed loans) are good steps in that direction. The more DeFi touches the “real” economy, the more sustainable it will be in the long run

Private credit is one of the fastest‑growing yield categories in DeFi. It offers higher yields: Maple’s current USDC APY is around 4.2% versus Aave’s 2%. Someone might note that 4.2% APY on what is essentially private credit versus 4.3% Treasury yields is a horrendous risk‑return trade, which it is. However, there are a few factors that make the future of on‑chain private credit potentially exciting: (a) democratization of access, (b) composability (use as collateral for loans and trading plus the ability to increase leverage via looping), and (c) capital formation for an underserved market.

I have covered a) and b) extensively before, however, capital formation for an underserved market is an interesting opportunity as well.

Skeptics would posit that private credit is currently a “borrower’s market,” with credit spreads tightening, a rise of covenant‑lite structures, and increased PIK usage. This means there is no rush for decent‑quality businesses to apply for loans sourced on‑chain if they can go to Apollo or a bank on similar terms.

However, there seems to be a gap in the small‑business loans market. In 2024, only around 15% of small‑business loan applications were approved by banks, and roughly 44% of small businesses did not even apply for these loans.

Hence, small businesses are increasingly shifting to fintech players. Ease of use, higher approval rates, and speed are some of the reasons.

This is where Maple and other private‑credit operators should and can play. For instance, USD.ai found a completely underserved market in GPU‑backed loans for emerging AI companies. Whether GPUs make for good collateral is a separate question, but USD.ai has made tens of millions of dollars in loans to dozens of counterparties. Peanuts in the grand scheme of things, but an interesting proof of concept.

Crypto has always served and created infrastructure for “outsiders.” In its current iteration, this means a user could be onboarded in Africa via a stablecoin wallet and then contribute directly to GPU‑backed loans in Silicon Valley, sidestepping Wall Street and the traditional financial system completely.

While trading has been the golden goose of crypto thus far, these “real economy” applications are more exciting. In the former, value concentrates from the many to the few; in the latter, it can be distributed from many to many.

However, for private credit to scale on DeFi rails, the operational math is challenging. Rough calculations suggest a firm needs 20–50 staff per 100 loans annually, capping these firms at roughly $1–5 billion AUM. Maple, for instance, currently has around 50 staff and approximately $4–4.5 billion AUM, implying a practical limit of about 100 loans per year. In other words, CeFi (centralized finance) does not scale in the same way that DeFi or even stablecoins do. However, there might be a solution closer than you think.

V. Transparency Gap in DeFi

While DeFi’s biggest advantage is its transparency, it still has a way to go.

Fully on‑chain DeFi arguably has great transparency (with some exceptions), but things break down when there is an off‑chain component in the flow. Overall, while DeFi was supposed to solve the transparency issue that would prevent another 2008‑type financial collapse, it has not yet done so.

For instance, you can see that a wallet received a loan from Maple. However, you cannot see who that wallet belongs to, what the loan terms are, whether covenants exist, or what the collateral actually consists of. We only know the collateralization ratio, which is a step in the right direction.

However, visibility ends exactly at the point where it would matter most. You have to trust that underwriters have done their due diligence.

The problem with private credit is both an incentive and information issue. Lenders and borrowers may benefit from lack of transparency: they can adjust distressed loans discreetly via hidden amendments, extensions, or PIK interest. There is a reason businesses choose more expensive private credit over cheaper high‑yield bonds for funding.

If the counterparty is a trading firm, the risk is that borrowed capital is repeatedly rehypothecated rather than used in a real underlying business, as the Elixir–Stream episode showed. Capital can be borrowed at one rate, redeployed into slightly higher‑yield venues, and then pledged or lent again, stacking apparent yield and intermediaries on the same collateral. By the end of that chain, there may be no genuine activity or asset underpinning the returns, only a fragile spread that vanishes when too many lenders try to redeem at once.

Maple’s dashboards are a bit more opaque, but there are also encouraging examples like USD.ai’s transparency dashboards. Their proof‑of‑reserves views show live collateral, active NVIDIA GPU loans at 11.5–15% yields, with individual loan status visible in real time. That is genuinely better than a quarterly filing. But due to the nature of private credit, these loans are still largely a black box.

Another promising solution is on‑chain rating agencies like Credora, but they still do not answer the question of who exactly the capital is going to.

The current iteration of fintechs and on‑chain lenders is the first serious attempt at serving the underserved lending market. Countless P2P lending firms have shut down over the years mainly due to liquidity issues and opaqueness. We should not repeat the same mistakes this time around. Transparency would be a step in the right direction.

VI. AI-Powered Credit

Private credit is a centralized endeavour with underwriters as the bottleneck to new deals and decisions. The fully decentralized DeFi version of this has not been feasible at meaningful scale for good reason. Technically, a DAO could receive debt applications and allocate based on token‑holder votes. The issue is low DAO voting participation and lack of domain expertise. More tokens do not equal more expertise.

AI will help in two ways: better packaging of investment decisions and improved information flow.

First, AI agents could process a large array of deals and give clear outputs to vote on, transforming pedantic things like loan structuring into things like default risk, industry risk levels, opportunity cost, and estimated risk‑adjusted returns. While agent development will still require a centralized dev shop, even this can be democratized using tools like Claude Code. The capital‑allocation “DNA” of a DAO can be codified into agent instructions. Those instructions can be voted on by token holders in proportion to their stake. While this does not solve turnout, it gives more understandable signals to vote on. Periodically, the agent’s “DNA” can be updated to reflect the changing token‑holder base.

Second, maybe more crucially, a layer of agents under a senior underwriter could negotiate loan terms, settle positions, service the loan book, collect interest, and manage the full credit lifecycle. The outcome is that a platform like Maple can manage multiples of its current volume without proportional headcount growth. The infrastructure to build this exists today.

Taken to its conclusion, this enables private‑credit risk models that update hourly and post on‑chain. Credit agents from lenders and borrowers interact directly, giving lenders real‑time visibility into counterparty health. The transparency gap is not a data issue but a reporting/effort issue. The data exists at the company level. It is a tooling and incentive problem, and AI agents are the most credible solution to it currently available. If anyone wants to build this, you are interested.

All of this shifts the math from opaque decisions like “should we lend to this counterparty,” which requires expertise most holders do not have, to “what kind of lender do we want to be,” which is a question any holder can engage with meaningfully, and where each DeFi protocol can differentiate itself.

VII. How DeFi Wins and Loses

The historical analogy worth keeping in mind is Visa in the 1970s. When Dee Hock built Visa, he was not selling a centralized financial product. He was selling an open acceptance network. The assurance that a card would work anywhere, with any merchant, without the issuing bank needing a direct relationship with each counterparty. The network’s value derived entirely from its openness. Exclusivity would have killed it.

DeFi’s structural equivalent is composability. The assurance that any asset, any protocol, and any yield strategy can plug into the same infrastructure without permission. That is the moat TradFi cannot replicate on permissioned chains, because permissioned chains require someone to grant permission, which means the network is only as open as whoever controls access decides it should be. Composability compounds in ways permissioned systems structurally cannot.

As discussed, DeFi wins by being first to genuinely new asset classes, not re‑tokenizing what TradFi already handles efficiently, but bringing net‑new asset classes on‑chain: GPU compute, reinsurance, DePIN networks, and AI agent economies, before traditional capital markets have the infrastructure to reach them. The first mover in a new asset class has a window before institutional capital arrives and compresses yields. That window is the opportunity. DeFi also wins through funding the long tail of borrowers and projects that traditional credit ignores, the original promise of banking the underbanked at scale.

DeFi loses if it becomes a yield‑extraction machine for Wall Street. The pattern is predictable and has precedent. DeFi innovates → Wall Street observes → Wall Street builds a compliant permissioned version → Wall Street pulls its capital into that version. JPMorgan’s Quorum in 2016–17 was the first attempt. It failed because institutions could not agree on governance, but corporate chains today may be running the same playbook, depending in part on whether they democratize validation (for example, Canton’s small validator set). DeFi’s institutional capital base migrates to a walled garden most cannot access. This is not constrained to DeFi either; neobanks like Robinhood are building their own prediction markets on off‑chain rails.

The second loss condition is more gradual. Competition for depositor capital forces protocols to offer higher rates to stay competitive. Higher rates require moving down the risk curve. While introducing new underserved asset classes on‑chain comes with inherent risks, a race to the bottom can cause DeFi protocols to become sloppy. Moving down the risk curve could eventually produce a blow‑up that resets confidence across the category. The Ethena situation was a relatively innocuous preview of this, Elixir less so.

The matrix that determines the outcome is yield against decentralization. High yield with high centralization is traditional private credit with blockchain branding, and TradFi wins that game on its own terms. Low yield with high decentralization is a composability platform without a business model. The only position that justifies DeFi’s existence as a distinct category is high yield with genuine decentralization, backed by the transparency infrastructure that makes that risk legible to the people taking it.

If you are working on any of these problems, you should reach out.

Work With Me

I advise protocols, funds, and financial institutions on DeFi market structure, tokenization strategy, and the intersection of AI and on-chain capital markets. If you are evaluating where to position in this space, building infrastructure that touches any of the dynamics described here, or trying to explain this market to institutional stakeholders who need a clearer frame, reach out.

LinkedIn: https://www.linkedin.com/in/jurgis-pocius/

Email: jurgis@evoaai.com

If this was useful, forward it to one person who should read it. If they are not already subscribed, the link is below. Every issue is an attempt to be early and specific about where the internet economy is actually going, not where the consensus says it is.

Read by teams at Solana Labs, Delphi Digital, Collab+Currency, 1kx, Multicoin Capital, and by operators at Prada, Dior, Nike, JPMorgan, McKinsey, HBS, and other leading funds and traders.

Subscribe now

The global personal luxury goods market is at an inflection point. After a decade of almost uninterrupted expansion, interrupted only briefly by COVID, the industry saw its first meaningful contraction in 15 years in 2024–2025, with the market stabilizing at around €358 billion in 2025 (flat to -2% year-on-year). Approximately 50 million luxury consumers have exited the market since 2022, effectively wiping out the aspirational buyer segment. Within this contraction, a brutal divergence has emerged: Hermès and Prada are gaining share, LVMH and Richemont remain resilient, and Kering is in structural crisis.

This report examines the seven houses that define the current luxury landscape: LVMH, Hermès, Chanel, Kering, Richemont, Moncler, and Prada, through the lens of financial performance, strategic positioning, and forward-looking catalysts. It is written for investors, operators, and strategists who want clear insight into where the industry is heading and which houses are positioned to lead.

Subscribe now

More than a Cyclical Contraction

The most alarming data point is not revenue decline; it is the erosion of the customer base. Between 2022 and 2025, the active luxury consumer pool shrank from roughly 400 million to about 330 million. This is not a cyclical pullback. It reflects:

1. Aggressive price inflation (2020–2023): Cumulative price increases of 40–60% across the industry have priced out aspirational buyers, likely permanently.

2. Value-perception collapse: More than half of surveyed luxury consumers say quality no longer justifies the price.

3. Generational shift: Gen Z advocacy for luxury brands continues to decline as spending rotates toward experiences, wellness, and cultural objects.

4. Resale cannibalization: A luxury resale market of roughly €35 billion, growing at 10–11% annually, offers a more accessible entry point and undercuts full-price demand.

   ---

Where We Are Now: Brand Performance

Hermès’ positive Q4 2025 earnings were no surprise to anyone following the sector, but the industry overall is in a period of growing pains driven by weaker Chinese demand and the disappearance of the aspirational consumer.

On public markets, Hermès and Richemont are the clear winners, the only major names whose share prices are in positive territory since January 2023.

In terms of scale, LVMH still dominates, accounting for roughly half of the large-cap luxury group universe by revenue. Kering, Chanel, Hermès, and Richemont each sit at around 20–25% of LVMH’s size.

Revenue growth tells the more interesting story. LVMH is slowing, while Hermès and Richemont continue to grow. Prada and Moncler follow. For the rest of the major houses, growth has decelerated sharply and, in many cases, turned negative.

Hermès’ operating margins (around 41%) tower above the rest of the sector. Richemont, by contrast, sits mid-pack despite strong top-line performance, in part because of a roughly 74% increase in gold prices over the past year, which weighs on margins for gold-intensive jewelry maisons.

What’s Working, What’s Not

Zooming into categories exposes the underlying pattern.

What’s doing well:

• Hermès’ leather goods have maintained desirability

• Miu Miu, now roughly 30% of Prada’s revenue, has become a magnet for younger consumers.

• Ultra-luxury jewelry: Richemont’s jewelry maisons (especially Cartier) and LVMH’s jewelry division led by Tiffany & Co. have outperformed, proving that true high-end, “wearable asset” products still command demand.

What’s not:

• Watches as speculative assets: The post-2021 speculative boom in watches has deflated, with prices correcting across the board. Yet even here, there is divergence: iconic models like the Cartier Tank have held value far better than the broader watch index.

• Overextended fashion brands: Gucci’s roughly 19% annual revenue decline recalls its overexpansion in the 1980s and underlines the danger of trend-chasing and logo fatigue.

  

  

  

Geographic Trends and the Asia-Pacific Myth

The common narrative of over-reliance on Asia-Pacific is false. In fact, two of the biggest growers this year (Hermès and Richemont) derive the majority of their revenue from that region, roughly 10 percentage points more than their peers.

Key geographic observations:

• Japan shows polarization: Richemont and Hermès gained ground while LVMH and Kering lost share

• United States demonstrates similar trends with clear winners and losers

• Kering is losing in all geographies

• Hermès and Prada grew across all geographies

• Richemont declined in its most important market despite overall strength

• The growth stories have largely been Japan and the Middle East.

• LVMH posted negative growth in all key markets

Flagship products with truly high-end luxurious status remain in demand, while product line extensions, speculative assets like watches, and aspirational goods have faltered.

Consumers have become skeptical. If secondary market prices drop or quality appears poor (paying for a logo rather than craftsmanship), they walk away. This mirrors the Gucci licensing debacle of the past.

Why the Luxury Model Is Failing

A review of recent financials and disclosures points to three structural failures across today’s major luxury houses, most visibly at LVMH and Kering.

1) Overextended Into Trends

Many brands have over-indexed on fast-moving trends at the expense of enduring icons and long-term brand equity. When hype becomes the primary growth driver, revenue volatility increases, product cycles shorten, and the perception of timelessness erodes. What was once cultural authority becomes cultural reactivity.

2) Erosion of Product and Service Quality

Luxury does not need to satisfy everyone, but it must be exceptional for the customers it chooses to serve. Aggregated analysis of customer reviews across major platforms consistently highlights the same weaknesses: slow responsiveness, uneven after-sales support, and inconsistent treatment across client tiers. The pattern is clear. Boutiques continue to sell a world-class promise, but many houses struggle to deliver world-class service at scale in a digital, omnichannel environment. This gap creates space for more agile, service-first competitors.

Two critical gaps emerge along the customer journey:

• The boutique experience anchor: High-prestige flagships function as atmospheric anchors for the brand. Performance is driven by high-touch hospitality, including bespoke events, cultivated sales-associate relationships, and carefully choreographed in-store experiences. This establishes a sentiment floor that sustains loyalty among top clients even when other touchpoints underperform.

• Service lifecycle misalignment: Post-purchase, many brands revert to defensive, policy-led service models. Returns, repairs, and complaints are handled through rigid, arbitration-style processes. This approach directly conflicts with the warmth and intimacy of the boutique experience, creating a persistent sentiment gap across the long-term ownership lifecycle.

  

  

3) Inflation and the Pullback of the Aspirational Buyer

Luxury price inflation has collided with a broader cost-of-living squeeze. Since 2020, aggressive ticket increases layered on top of general inflation have pushed aspirational buyers, and even parts of the affluent cohort, to pull back. Industry analysis increasingly links the loss of tens of millions of customers since 2022 to a combined effect of price elevation, value-perception collapse, and macro pressure. The active luxury customer base is shrinking, with the mid-tier absorbing the majority of the impact.

How Luxury Is Trying to Recover

In response to the shrinking customer base and weakening mid-tier demand, major groups are pivoting their strategies rather than standing still.

1) The Hard-Luxury Pivot

Jewelry has become the strategic priority for almost every group, viewed as a resilient “wearable asset” that outperforms soft luxury (ready-to-wear) in periods of volatility.

• LVMH is aggressively elevating its Watches & Jewelry division, with Tiffany & Co. pursuing an upmarket “elevation strategy” (renovated flagships, high jewelry collections like Blue Book) and Louis Vuitton expanding in haute horlogerie and high jewelry through collections such as Awakened Hands, Awakened Minds. Bvlgari has expanded its Valenza manufacturing site to support this jewelry growth.

• Kering is pursuing an industrial development strategy in jewelry, including a staged acquisition of Raselli Franco Group to secure manufacturing capabilities and tighter control of the value chain for houses like Boucheron and Qeelin.

• Richemont, already the category leader, continues to double down on its jewelry maisons (Cartier, Van Cleef & Arpels), which are growing at double-digit rates and remain its primary growth engine, far outpacing the watch division

  

2) Expansion Into Beauty, Fragrance, and Hospitality

Alongside hard luxury, groups are leaning into fragrance, skincare, color cosmetics, and hospitality to keep their brands present in customers’ daily lives while offering lower-priced, higher-frequency entry points. These moves let them monetize brand equity even as big-ticket fashion purchases slow.

3) Return to Direct-to-Consumer Control

Most major houses are steadily pulling distribution back to DTC channels. The rationale is straightforward: tighter pricing control (no discounting or grey promotions), access to rich first-party customer data in an AI-driven world, and full control over how the brand is staged at every touchpoint.

4) Experiences as the new moat

Finally, brands are investing more in experiences, from flagship “world-building” stores and private events to travel, culture, and hospitality, to make entry into their universe feel like stepping into a different world rather than just buying a logo. The open question for the next cycle is whose world will feel compelling enough to justify the price of admission.

Luxury is not dying, but the old model of scaled aspiration is. The future belongs to houses that protect their mystique, treat culture and service as core products, and build worlds that a smaller, more discerning audience is desperate to enter. From here, we can look at the issues through each brand’s lens.

Inside the Luxury Houses

Hermès: Leader in a Stagnant Market Real Physical Luxury

Hermès is the clear outlier in a stagnant luxury market. In 2025, revenue crossed €16 billion and recurring operating margins reached 41%, with the market valuing the house at a PE of around 49 versus roughly 20–25 for LVMH. That valuation gap only makes sense if Hermès is doing something fundamentally different from the rest of the industry. And it is.

1. Slow Luxury as the Ultimate Flywheel

Hermès’ core engine is leather goods and saddlery, still growing in the low teens even as peers stall, with sales up about 12–13% in 2025 on the back of new lines like Arçon, Faubourg Express, and Médor. Each Birkin is made by a single artisan over roughly 18–24 hours; there is no assembly line, which structurally caps supply at an estimated ~100,000 bags per year versus roughly 400,000 for a similarly sized competitor like Louis Vuitton. This enforced scarcity means every bag has a buyer, there is no need for discounting, and resale prices stay supported.

The result is that Hermès bags have behaved more like an asset class than a fashion accessory. Estimates put the average appreciation of Birkin/Togo bags at around 14% annually over long periods, with a Birkin 35 in Togo leather now changing hands at roughly €11,600 and pristine models commonly reselling at 2x retail. In practice, Hermès cannot sell dramatically more bags without breaking the model, so incremental profit growth increasingly has to come from cross-selling into adjacent categories (jewelry, homeware, ready-to-wear, and high-end travel goods) to the same ultra price-insensitive client.

2. Cultural Events Over Hype-Driven Marketing

Where most luxury houses spend heavily to rent cultural relevance from celebrities, Hermès has built a system that makes that unnecessary. In 2025, LVMH spent around 37% of revenue on marketing and selling, whereas Hermès spends closer to 4%, yet Hermès still commands the higher multiple and stronger pricing power. The house has no celebrity brand ambassadors, no star “creative director” fronting the brand, and no dependence on hype cycles or influencer-driven trends.

Instead, Hermès invests in cultural events that are structurally on-brand. “Hermès in the Making” lets the public meet artisans and see how materials are transformed, while “Mystery at the Grooms’” stages an interactive, free narrative around the company’s 16 métiers in cities like New York and Tokyo. These are live, traveling proofs of concept for the brand’s core proposition. You aren’t buying stuff, you’re buying craft, time, and story.​

3. Category- and Region-Specific Storytelling

Hermès treats fashion and product launches as chaptered narratives rather than isolated runway moments. After the main Paris show, it brings “Chapter 2” of collections to key markets like Shanghai, tailoring silhouettes (elevated streetwear, layered outerwear) to local aesthetics without diluting the brand’s equestrian and artisanal core.

At the same time, it builds category-specific stages:​

• High jewelry: Exclusive events like “Les formes de la couleur” in Singapore and Tokyo.

  

• Home: Presence at Salone del Mobile in Milan to position Hermès as a serious design house, not a merch afterthought.

• Watches: Launched new line at Watches and Wonders Geneva, even as the broader Swiss watch industry cools.

Despite venturing abroad, Hermès stays true to its French and equestrian roots. Since 2011, it has organized Saut Hermès, an annual three-day showjumping competition featuring elite equestrian action.

4. Vertical Integration as a Moat

Hermès’ business model is designed around one idea: control everything that matters to craft and brand, from the farm to the flagship.

In leather goods and saddlery, more than half of production now comes from company-owned or exclusive workshops, with the 24th leather facility opening in L’Isle-d’Espagnac in 2025 and several more slated through 2030. The group also directly manages tanning, textile printing, and parts of watchmaking, expanding its Noirmont site in Switzerland to boost in-house movements and cases.

Hermès has also taken minority stakes in suppliers like the 15% in Lanificio Colombo, an Italian mill known for cashmere and vicuña that has long supplied its shawls and knitwear, securing long-term access to rare fibers without absorbing the family business.

It has rolled out laser marking on hides to trace leather back to farm level extending and is phasing out fossil fuels in new workshops in favor of electrification and renewables.​

Vertical integration helps Hermés guarantee quality, scarcity, and supply-chain sovereignty in a world where every other brand is fighting over the same outsourced capacity. Its reliance on local suppliers (74% of production operations in France, 98% of its materials from European suppliers) also makes it easier to navigate the volatile geopolitical and tariff issues.

5. Holding Structure That Is Built for the Long Game

Most public luxury companies are structurally set up to chase quarterly EPS and appease activist investors. Hermès is engineered to ignore that short-termism. The H51 family holding structure and the dominance of the Hermès heirs mean that listed shares are largely non-controlling, which makes hostile takeovers or short-term strategic pivots nearly impossible.

Internally, Hermès reinforces this long-termism with how it treats its workforce. After another year of record results, the company awarded thousands of employees a meaningful profit-sharing bonus (€3,000 per employee in 2025), on top of salary, framing artisans and sales staff as equity-like stakeholders in the brand’s success. That incentive design, layered on top of an industrial system that bans fossil fuels in new sites and wins transparency awards for climate reporting, creates a culture where doing things that “don’t scale” (like one artisan per bag, handsewn saddlestitching that machines can’t replicate) is not a bug but the point.

A Profitable Golden Cage

Hermès is not merely selling handbags, shoes, or scarves, but a scarce, culturally loaded object whose utility is part art, part social signal, and part store of value.

This pure heritage positioning has another advantage in that it effectively makes the company “AI-proof,” for better or for worse. Mechanizing manufacturing or over-optimizing advertising would, if anything, dilute the brand’s mystique rather than strengthen it. Instead, Hermès can introduce AI slowly and invisibly in back-office and analytical workflows, while keeping the client-facing experience resolutely human and artisanal. Many other luxury brands, more dependent on high-volume marketing and with messier internal organizations, stand to benefit more visibly from deploying the technology.

In downturns, that combination of extreme pricing power, tightly controlled vertical production, and a corporate structure insulated from short-term shareholder pressure is exactly what you want to own. While peers cut prices, chase trends, and overextend into wholesale and outlet channels, Hermès can simply keep doing less, better.

Key Risks

Hermès still relies heavily on real leather and exotic skins, and at some point, especially with younger generations, this may become less desirable as sustainability concerns and animal advocacy gain momentum. While Hermès has outperformed many competitors, its core growth engine in Asia-Pacific is facing softer demand and increasingly confident local luxury brands.

Given constrained production capacity, its main revenue lever, price, inherently has limits; beyond a certain point, customers may shift to other houses or adjacent categories. Growth has also been uneven across newer segments, for example, categories like watches have seen more modest traction, in part because Hermès is still primarily known for its leather goods.

In some ways, Hermès has built itself a golden cage: the very scarcity and artisanal constraints that protect the brand also cap its degrees of freedom. But as long as it maintains a category-of-one position in ultra-luxury leather goods, it is likely to remain a very profitable one.

LVMH: A “Recession-Proof” Private Equity Platform

For a deeper dive on LVMH’s strategy, I previously wrote about it here, here, and here.

A PE Fund in Luxury Company’s Clothing

LVMH is essentially a private equity fund in a luxury company’s clothing. Unlike competitors such as Hermès, which emphasize strict organic growth and holding assets indefinitely, LVMH actively prunes and expands its portfolio to optimize returns, much like a PE fund.

It ruthlessly cuts business lines when they are no longer relevant, for instance the disposal of Off-White in September 2024 and the sale of a significant portion of DFS (Duty Free Shoppers) in January 2026 to refocus on more profitable assets. LVMH also doubles down on winners, raising its stake in Loro Piana to 94% by acquiring an additional 9% from minority shareholders for €1.0 billion in July 2025. It seeks to consolidate control of the luxury market as a whole, successfully acquiring players like Tiffany & Co. and unsuccessfully attempting to take over Hermès.

More recently, LVMH entered a strategic agreement in September 2024 to acquire up to a 22% stake in Double R, the holding company controlled by Moncler CEO Remo Ruffini. This indirect structure allows LVMH to back a “winner” and gain exposure to Moncler’s growth while keeping the founder firmly in charge, a classic PE-style growth equity playbook.

LVMH is also quietly centralizing its supply chain and technology stack, using agentic AI and digital supply chain twins to model factories and inventory in real time, with the goal of boosting efficiency, widening margins, and stripping out redundancies.

This strategy is well-suited to fast growth and internal hedging. Strength in Jewelry and Beauty has offset softness in Fashion & Leather and Wines & Spirits, with high jewelry (e.g., Tiffany’s 2025 Blue Book “Sea of Wonder”) and new beauty launches such as La Beauté Louis Vuitton acting as incremental growth engines. Sephora, an atypically “mass” brand in the LVMH portfolio, can be read as a feeder: it pulls in younger and more price-sensitive customers at the top of the funnel, then funnels a subset into more premium, higher-margin offerings.

Culturally Relevant Experiences as the Demand Driver

Unlike Hermès, LVMH relies heavily on celebrity ambassadors, acquisitions, and cultural events to maintain relevance. You can think of it as a solid heritage base that stays current by associating with culturally timely brands and personalities. This is both a blessing and a curse: it is expensive, and audiences are fickle. Cultural trends can turn quickly, as we saw with hype cycles around names like Supreme and CryptoPunks.

On the other hand, collaborations allow LVMH to borrow cultural relevance without tying itself too deeply to any single underlying brand. If a brand or celebrity loses momentum, LVMH can simply let the partnership lapse and pivot to the next one.

The next decade in luxury will be increasingly defined by experiential offerings. Consumers have shifted toward valuing experiences over products, and luxury brands have been quick to follow. LVMH is pushing this to the next level.

While the losses from this category are substantial (recurring operating losses of roughly €491 million in “Other activities and eliminations” in 2024), it is a core part of brand building and long-term desirability. Experiential highlights include:

• LVMH as an Official Partner of Formula 1 under a landmark multi-year global partnership.

• Cheval Blanc and Belmond, its luxury hotel brands, with locations in Paris, the French Alps, St-Tropez, the Maldives, the Caribbean, and upcoming properties in Dubai (2029) and Porto Cervo, Italy (2026).​

• A strategic partnership with Accor to revive and develop the historic Orient Express brand.​

• Ownership of Royal Van Lent, a high-end yacht builder, rounding out the ultra-luxury travel and leisure ecosystem.​

Even with a roughly $100 million annual Formula 1 deal and a strong presence at the Paris Olympics, there is still a noticeable gap in the sports/lifestyle vertical that LVMH could fill.

From the outside, two experiential gaps stand out:

• Lifestyle/high-end sports gap
  LVMH could begin investing in unique experiential and real estate assets like members-only tennis clubs. Tennis, in particular, is seeing renewed interest among millennials and younger cohorts; it is simultaneously a lifestyle, an experience, and a real estate play. More broadly, there is room to build around the “top 10 elitist sports” cluster: polo, yacht racing/sailing, equestrian, golf, fencing, rowing, tennis, alpine skiing, squash, and Formula 1, as branded ecosystems rather than just sponsorships.

• 1:1 hospitality offerings
  Beyond existing hospitality investments such as Cheval Blanc and Belmond, there is an opportunity in ultra-niche travel destinations for VIP clients, think Lapland, a popular choice among Asia-Pacific consumers and celebrities. Imagine ultra-high-end, fully serviced properties owned by LVMH, effectively “black-card Airbnb” experiences gifted or reserved for top clients only. Standard luxury experiences will not be enough; they will need to be luxury experiences that no one else can access.

Recent examples of this “elevated experiential” strategy include the reopened Milan Louis Vuitton flagship and immersive concepts like the Blue Box Café in New York, which combine retail, dining, and cultural programming. The Da Vittorio Café Louis Vuitton and DaV by Da Vittorio restaurant meld simplicity with refinement, pairing “luxury snacking” with three‑Michelin‑star cuisine and reimagined classics like osso buco with saffron risotto served in the shape of the Louis Vuitton Monogram flower. These openings sit within a broader elevation strategy at Tiffany & Co., which has renovated nearly a third of its network to enhance brand desirability and the quality of the customer experience.

The risk is that LVMH cannot afford to get sloppy. Acquisition for acquisition’s sake will not work in a slowing luxury cycle. The group also reported around €1.065 billion in losses from exchange rate fluctuations, which raises uncomfortable questions about why FX hedging was not more effective given its scale, geographic footprint, and relatively predictable currency exposures. For an operator this sophisticated, failing to better mitigate known FX risk feels like an unforced error.

PRADA: The Miu Miu Success Story

Prada Group enters 2026 in an unusually strong position: €5.4bn in revenue (+17%), €1.28bn in EBIT (+21%), a 23.6% EBIT margin, and net cash of around €600m after four straight years of double-digit like-for-like growth.

Structural Challenges

The group is now a two-speed machine: Miu Miu in hypergrowth versus a Prada “mothership” that is solid but slower (low single-digit growth or slight decline, depending on the period). If Miu Miu’s aesthetic moment cools, as fashion moments always do, the group needs Prada to carry more weight than it currently does.

A second issue is self‑inflicted: the Versace acquisition adds complexity and near-term margin drag. Prada is giving up a near-pristine balance sheet and taking on real execution risk.

Strategic Response

First, Prada is keeping Miu Miu’s engine running. Miu Miu is being treated as the group’s primary growth engine. Its “schoolgirl chaos” aesthetic (micro‑minis, ballet flats, shrunken knits, exposed lingerie) has become a dominant cultural reference, driving extraordinary retail momentum. The brand maintains the momentum via cultural programming: Women’s Tales films, Literary Clubs, Atheneum pop‑ups, and collaborations like New Balance × Coco Gauff. A new series of pop‑up stores across Asia this spring is designed as fully immersive spaces that pull consumers deeper into the Miu Miu universe. On top of that, the L’Oréal fragrance launch (“Miutine”) pushes Miu Miu into beauty and adds a high‑margin royalty stream to the growth engine.

Second, Prada is raising the “mothership” growth rate without breaking it. The Prada brand is being managed for durable desirability revitalising icons like the Galleria bag, adding new leather goods lines (Explore, Etude, Dada), and using cultural platforms (Prada Mode London, Art Basel activations such as “30 Blizzards”) to stay intellectually relevant rather than chasing viral trend cycles.

Third, Prada is using Versace as its big portfolio bet. It bought the brand at a distressed price (around €1.25bn EV, funded with roughly €1.5bn of new debt) and is attempting a clean identity reset. Donatella has stepped down and Dario Vitale (formerly design lead at Miu Miu) is in place to rebuild product and relevance, with early buyer reception reported as strong. Governance signals long-term intent with Lorenzo Bertelli installed as Versace’s chair. Near term, Versace is likely to drag margins, while medium term the upside is a non-overlapping customer base and a stronger Americas mix.

Fourth, Prada is doubling down on brand control. It is using its own stores and in‑house production to defend full-price selling, manage distribution tightly, and build a tight loop between physical retail and digital touchpoints where events and hospitality create emotional pull, and directly operated stores capture both the transaction and the data.

Prada’s ultimate edge is control, built on three pillars:

• Around 90% of revenue comes from directly operated stores, enabling pricing discipline, control of the brand environment, and customer‑data ownership with minimal wholesale leakage.

• 26 owned factories provide vertical integration that protects quality, speed, and supply stability, and create a clear integration path for Versace manufacturing.

• A historically strong balance sheet (net cash pre‑deal) gives the group room to absorb a turnaround without starving its core brands.

This is why comparisons to “Italy’s LVMH” keep surfacing, but the model is narrower and more focused: a small portfolio of distinct identities sharing infrastructure, not creative voice.

Richemont: Narrow, Hard-Luxury Powerhouse

Richemont, owner of Cartier, Van Cleef & Arpels and Buccellati, enters 2026 as the hard‑luxury benchmark. Jewelry and watches are carrying the group through a messy macro environment, with growth still solid even as cost inflation and weaker China demand drag on performance. The core tension is that the portfolio is compounding, but each incremental euro has to work harder against higher gold prices, a strong Swiss franc and patchy demand across Asia.

What makes Richemont interesting is how well its structure fits this moment. Jewelry is the engine. Cartier, Van Cleef & Arpels and Buccellati, now joined by Vhernier, have become an industrial compounding machine, with high single‑ to mid‑teens growth at constant FX and operating margins in the low‑30s even after the spike in gold. The playbook is deliberately conservative: rely on “timeless” platforms such as Panthère and Alhambra, then punctuate them with jewelry events and one‑off pieces.

Presenting these collections in the right way is key. Cartier stages invite‑only En Équilibre presentations, Van Cleef & Arpels builds full worlds around collections like l’Île au Trésor, Buccellati hosts intimate gatherings in Italy, and Vacheron Constantin, on the watch side, places an elaborate piece such as La Quête du Temps in the Louvre, turning it into a culturally situated object where history, craftsmanship and engineering are read together. That mix of timeless product and well-curated cultural experience is precisely what very wealthy clients are paying for.

Richemont has also rebuilt its operating model around control. Roughly three‑quarters of revenue now comes from its own boutiques and e‑commerce, with retail growth outpacing wholesale. That gives the group much tighter grip on pricing, assortment and client experience than wholesale‑heavy models that peers are still trying to unwind. Regionally, it is leaning into depth where demand is resilient, focusing on local clients in the US, Japan and the Middle East, while treating Greater China as a market to manage rather than a growth crutch and accepting that recovery there will be gradual and uneven. Underneath sits a fortress balance sheet. Net cash at year‑end 2025 is in the mid‑single‑digit billions, and Richemont has used that position to keep Capex elevated in jewelry manufacturing and flagship boutiques, acquire assets such as Vhernier, fund the clean exit from YNAP, and still pay dividends and buybacks while remaining in net cash.

Structural Challenges

Richemont’s main operational concern is not Cartier’s jewelry watches, but its standalone Specialist Watchmakers division, which includes IWC, Panerai and Vacheron Constantin. A strong franc, higher precious‑metal costs and softer volumes are moving through a largely fixed‑cost base and squeezing profitability, which in turn weighs on group margins.

At the same time, the long tail of fashion and accessories, including Chloé, Dunhill and Alaïa, remains loss-making, even if those losses are narrowing. That reinforces the idea that Richemont’s natural home is hard, not soft, luxury.

Strategic Response

Richemont’s response is unusually disciplined for a listed group. Rather than chasing volume, it is leaning on the equity of its maisons to implement “balanced and targeted” price increases and is explicitly prioritizing unit economics over headline growth.

On the cost and capacity side, it has resisted the reflex to cut capex. Instead, it is pushing more investment into jewelry manufacturing through new workshops, expanded lines, craft training and selective flagship upgrades in key cities, keeping the retail network high‑quality and curated rather than ubiquitous.

The exit from Yoox Net‑a‑Porter in 2025, via a sale into Mytheresa’s LuxExperience platform, a clean deconsolidation and a one‑off hit from discontinued operations, is part of the same functional discipline. Hard luxury has been put back at the centre, platform risk has been externalized and the P&L has been de‑cluttered.

Heading into 2026 and beyond, Richemont’s thesis is intentionally narrow. Instead of trying to match LVMH’s breadth or Kering’s fashion‑turnaround story, it is effectively betting that controlling the leading jewelry maisons, a tightly curated set of prestige watch brands and most of its own distribution is enough to ride out a volatile decade.

Chanel: A Surprisingly Innovative Heritage House

Chanel enters this slowdown as one of the few truly global, fully independent mega-brands in luxury. Ready-to-wear, leather goods, watches, jewelry, and beauty all sit under a privately held, debt-free structure, which gives the house more freedom than listed peers to invest through the cycle rather than defend quarterly margins.

Structural Challenges

Chanel is navigating a cyclical slowdown from a position of strength. After years of aggressive pricing, demand is softening among aspirational clients. In 2024, operating margins fell to roughly 24% from the low-30s, with operating profit down about 30%, even though the brand still generates strong cash.

The issue is not brand desirability, but rising expectations. As customers buy fewer pieces, every product, interaction, and experience has to feel earned. Luxury is no longer forgiven for friction, whether digital, physical, or experiential.

Chanel’s structure gives it room to respond differently. The house is privately held, debt-free, and insulated from quarterly market pressure. Instead of defending margins, 2024 became a year of higher investment (Chanel’s latest reporting period), with a clear step-up in capex and spend on brand support, real estate, and craft.

Headcount rose from about 36,500 to more than 38,400, even as peers froze hiring. Chanel has committed to net-zero emissions by 2040, sourced 99% of its electricity from renewables in 2024 with a target of 100% in 2025, and continued upgrading boutiques to higher environmental standards. The signal is long-term intent, not short-term optimization.

Strategic Response

Chanel’s strategy is focused and selective.

First, investing through the downturn. Capex rose sharply as the house expanded boutiques and flagships, including a Watches & Fine Jewelry store on New York’s Fifth Avenue and CHANEL & moi – Les Ateliers in Seoul, and scaled CHANEL & moi care and repair services globally. Those Ateliers focus on care, repair, and longevity, reinforcing the idea that Chanel pieces are maintained, not just sold.

Second, using AI where it actually moves the needle. Chanel is using AI to remove purchase friction and win attention in crowded feeds:

• Lipscanner (beauty shade-matching + AR try-on): +32% lipstick conversion and –27% return requests tied to shade mismatch; reported 1.2M+ downloads and strong app-store ratings, turning color uncertainty into a unique advantage.

• Storyteller Nail (generative creative at scale): generated 45,000 shade-specific micro-stories; nearly tripled view duration (to ~8 seconds), lifted click-through, and increased add-to-cart, making ads feel personal without diluting brand control.

• Conversational assistants (service + internal copilot pilots): handled a majority of routine inquiries end-to-end, improved first-contact resolution, reduced handle time, and increased CSAT; also reclaimed hours weekly for boutique teams by reducing manual lookups. (The strategic point: AI protects the “impeccable service” promise while scaling.)

Third, making physical retail culturally relevant again. Chanel is staging specific, unflattenable moments rather than generic shows, from the Métiers d’art collection in Hangzhou, inspired by coromandel screens, to Cruise repeat shows in Hong Kong designed to create local cultural heat, not just global impressions.

Fourth, expansion with intent. In 2025–2026, Chanel is expanding in India and Mexico and adding boutiques across Mainland China, Japan, and Canada, positioning stores as destinations as much as transactional spaces.

Fifth, securing savoir-faire and innovation. Chanel continues to invest in innovation infrastructure, partnerships with institutions such as the MIT Media Lab, internal entrepreneurship programs, and selective external stakes. Its 25% investment in independent watchmaker MB&F, following earlier stakes in Romain Gauthier and F.P. Journe, is about protecting high-end watchmaking capabilities and keeping access to rare skills inside the ecosystem.

Kering: A Prime Example of What Can Go Wrong

Kering enters 2026 as the laggard among the global luxury groups. Once defined by Gucci’s outperformance, it now faces falling sales, collapsing profits, and a portfolio whose desirability has slipped just as the market has become more polarized between true ultra‑luxury and everyone else.

Structural Challenges

Kering is a prime example of what can go wrong in modern luxury when brand equity and positioning slip. Group revenue fell 13% in 2025, and net income attributable to the group dropped more than 90%, far worse than the broader sector, even though it faces the same Asia–Pacific cooling as peers.

The core problem is brand desirability across the portfolio. Gucci has lost momentum with core and aspirational clients. Alexander McQueen is in restructuring and has seen a “sharp decline in footfall” and Balenciaga is no longer generating previous levels of interest, with Kering explicitly stating that it needs to revive interest in the house and acknowledging that the prior creative direction did not capture demand in a more polarized, uncertain environment.

Kering is also over‑indexed to aspirational rather than ultra‑high‑end clients, the segment that has cooled the most, and has historically relied more on wholesale than the likes of Hermès, LVMH or Richemont. Gucci deliberately cut wholesale revenue by 34% in 2025 to limit access to only the most exclusive accounts, accepting a near‑term revenue hit to try to rebuild scarcity and elevate the brand. At the same time, the broader luxury market is shifting incremental innovation from pure product to experiences, hospitality, dining, travel, events, where entertainment, emotion and ethics matter as much as the object itself, and Kering has been slower to build that layer at scale.

Kering is over-indexed to aspirational (vs ultra-luxury) clients, the segment that has cooled the most, and has relied more heavily on wholesale than peers. Gucci deliberately cut wholesale revenue by 34% in 2025 to limit access to only the “most exclusive accounts,” accepting short-term revenue loss to elevate the brand. In parallel, the wider luxury market is shifting innovation from pure products to “experiences” (hospitality, cruises, fine dining), emphasizing entertainment, emotions, and ethics over goods alone.

Strategic Response

In response, Kering is executing a multi-pronged transformation it describes as “flawless execution” of a new roadmap.

First, creative shock therapy at the flagship brands. Kering is attempting to reset its two most important fashion houses with high‑profile creative appointments. Demna, formerly synonymous with Balenciaga, became Gucci’s Artistic Director in March 2025, and his debut “La Famiglia” collection has been credited internally with sparking renewed interest in the brand. In parallel, Pierpaolo Piccioli, ex‑Valentino, was appointed Artistic Director of Balenciaga in May 2025 to steer the house back toward growth and away from a more polarizing phase.

Second, a governance and leadership overhaul. In September 2025, Kering split the roles of Chair and CEO, appointing Luca de Meo as Group CEO while François‑Henri Pinault stayed on as Chair. One of de Meo’s first moves was to put Francesca Bellettini, Kering’s long‑time Deputy CEO and head of Saint Laurent, in charge of Gucci as President and CEO, signaling that the group’s most important asset will now be run by one of its most operationally minded leaders.

Third, a strategic pivot in beauty. In October 2025, Kering agreed a long‑term strategic partnership with L’Oréal and sold Kering Beauté, including Creed, for roughly €4 billion. In exchange, its brands secured multi‑decade beauty licenses, effectively swapping in‑house execution risk for a high‑margin royalty stream and immediate balance sheet repair.

Fourth, asset management and vertical integration. Kering has been reshaping its asset base at both ends of the value chain. It sold stakes in prime properties on Paris’s rue de Sèvres and New York’s Fifth Avenue to Ardian while retaining a significant minority interest, freeing up capital without fully exiting strategic locations. On the production side, it has moved to secure more of the value chain in hard luxury, signing an agreement to acquire Raselli Franco Group, one of Europe’s leading independent jewelry manufacturers, and buying eyewear suppliers such as Visard and Lenti to strengthen Kering Eyewear’s integrated model.

Fifth, operational rationalization and channel control. Gucci closed a net 25 stores in 2025 and is shutting outlet locations to focus on “highly exclusive venues.” At group level, wholesale revenue fell nearly 20% as Kering aggressively cut third‑party retailers to push more business into its own network, again choosing near‑term pain in service of tighter brand and pricing control.

As the group enters 2026, management is framing this as the start of a multi‑year reset: a leaner, faster Kering that has to rebuild brand desirability, regain control of distribution and mix, and gradually rebuild margins and cash generation rather than relying on the easy growth of the last cycle.

Moncler: Between Resilient and Stuck

Moncler enters 2026 as a solid but not spectacular performer in luxury. It has largely held revenues flat while others have swung from boom to correction, but it has not broken out the way true ultra‑luxury winners have, and its growth engine is starting to sputter.

Structural Challenges

Moncler Group sits in the middle tier of luxury performers heading into 2026. It has avoided the sharp revenue and profit declines seen at players like Kering, but it is not matching the structural momentum of houses such as Hermès or Richemont.

The immediate issue is stagnating growth. For the first nine months of 2025, group revenue was down about 1% reported and essentially flat at constant exchange rates. Two things explain most of that stall. Tourism weakness hit EMEA and Japan, weighing on the Moncler brand, which still over‑indexes to traveling luxury customers. At the same time, the group is deliberately sacrificing wholesale revenue as it cleans up distribution and elevates positioning. Moncler brand wholesale fell around 5% at constant FX, with Stone Island down more, creating a short‑term drag on the top line.

Beneath the numbers, there is a more structural question. Moncler’s technical innovation has, to some degree, become normalized. As high‑end outdoor brands such as Arc’teryx expand consumers’ choices in performance outerwear, some customers are asking whether €1,000‑plus price points for Moncler are still fully justified. At the same time, the brand has edged close to hype saturation, a place that has historically been dangerous for long‑term equity, as seen in prior cycles at street and “hypebeast” names like Supreme or NFT‑driven drops. The recent “Warmer Together” campaign, built around heritage pieces like the Maya 70 and Bretagne with a message that “warmth was never about the outside”, is a conscious pivot back toward function, warmth and human connection over pure hype.

Structural Positioning

Moncler’s resilience is rooted more in structure than in growth. The group plays in “luxury performance,” a category that tends to hold up better than pure fashion when sentiment tightens, because cold‑weather function is a real need as well as a status signal. Nearly 90% of Moncler brand sales come from direct‑to‑consumer channels, and the group reports gross margins above 78%, giving it far more control over pricing, experience and data than wholesale‑heavy peers. Net cash of roughly €1.3 billion gives Moncler the balance sheet to accept short‑term revenue pressure without compromising strategic choices.

The two‑brand architecture adds another layer of protection. Moncler and Stone Island speak to different audiences, global luxury performance versus technical street culture, which allows the group to manage different cycles and aesthetics without over‑relying on a single “it” moment.

Strategic Response

Moncler’s response is focused rather than reactive.

First, it is accelerating DTC and accepting near‑term pain. Moncler is continuing to push more of the business into its own stores and digital channels and to prune wholesale aggressively in both brands. That weighs on reported growth, but it tightens control over price, assortment and brand environment. The strategy is already showing up in the Americas, which outperformed other regions in the latest period, driven by solid direct‑to‑consumer demand in the US.

Second, shifting from hype to community obsession. The group is trying to reframe Moncler around belonging and lived experience rather than constant drops. The Moncler Grenoble Fall/Winter 2026 show in Aspen, staged as a nighttime snowmobile ride through the forest to a snow‑covered mountain runway, is a blueprint: a physically exclusive event for the brand’s global community that generates durable cultural value and organic digital reach, instead of a one‑off stunt.

Third, re‑anchoring the brand in purpose and performance. The aforementioned, “Warmer Together” marks a deliberate move away from pure hype dynamics and back toward emotional relevance and functional credibility, reminding customers that Moncler earned its pricing power by keeping people warm in extreme conditions as much as by being seen in the right places

In short, Moncler’s problem is flat growth and more volatile tourism flows at a time when its once‑distinct technical edge feels less unique. It is countering that with a mix of DTC acceleration, tighter distribution, and extreme experiential marketing that leans into community, heritage and performance rather than chasing the next hype cycle.

The Forces That Will Matter in Luxury

Sustainability May Be Overrated

Over the past decade, luxury has treated sustainability as both a moral obligation and a competitive superpower. Heading into 2026, it looks more like regulated hygiene than a real growth driver. The brands winning on revenue and market cap are not the ESG darlings, and looming rules like the EU’s Digital Product Passport are turning “green” from a differentiator into a simple licence to operate.

Regulation is now closing in. The EU’s Digital Product Passport (DPP) regime, expected to be mandatory by 2030, will require brands to assign each product a scannable digital ID capturing origin, materials, manufacturing and recyclability. That will lift traceability expectations and compliance costs, but it also pushes sustainability further into the realm of licence‑to‑operate rather than a source of incremental growth.

The market data tells a similar story. Across major listed luxury names, there is no meaningful positive correlation between ESG scores and either revenue growth or share‑price performance; when you plot LSEG and S&P ESG scores against 1‑ and 3‑year returns, you see something closer to the opposite, with lower‑scoring ESG names often outperforming over three years. Hermès, Prada and Richemont, three of the strongest stock and revenue performers of the last few years, sit near the bottom of the ESG rankings table in 2025.

The same pattern holds when you run the numbers against revenue growth, no statistically significant positive relationship, and a visible negative trend line where higher ESG scores often map to weaker top‑line performance. Consumers may say sustainability matters at the checkout, but in aggregate they still reward desirability, scarcity, and execution first.

Against that backdrop, the sustainability strategy of the major houses going into 2026 looks less like a commercial bet and more like regulatory and operational inevitability. LVMH’s LIFE 360 roadmap is a good example: by 2026 it targets zero virgin fossil‑based plastic in customer packaging, a 50% cut in energy‑related emissions versus 2019, 100% renewable or low‑carbon energy for stores and sites, and full certification of “strategic” raw materials, with all new products carrying a customer information system, effectively a proto‑DPP layer, and all staff trained on environmental issues via the LIFE Academy.

Hermès is taking an even more supply‑chain‑centric route, aiming by 2026 for all raw materials to be deforestation‑ and conversion‑free, with strict policies for leather, wood, rubber, palm and soy and a 2020 cut‑off date baked into supplier contracts, backed by certifications such as FSC and LWG. Chanel, meanwhile, is pushing industrial circularity: by 2026 it plans for roughly 30% of handbags and 50% of shoes to use recycled structural components, has launched “Nevold” to recycle cotton, wool, silk and cashmere at scale, and is working toward eliminating plastic components from shoes and bags entirely.

At the industry level, 2026 is essentially a dress rehearsal for mandatory DPPs. Groups like Kering, Moncler and Richemont are building digital product registries to comply with the EU’s Ecodesign for Sustainable Products Regulation, which will require each item to carry a verified data trail on origin, composition and recyclability; Moncler, for example, has already moved beyond 50% recycled nylon by 2024 and is formalising “thinking circular” as a core operating principle, with the critical work happening in PLM, ERP and traceability architecture rather than in slogans.

Where sustainability may actually start to matter, commercially, is not in first‑sale demand but in supply chain economics and secondary-market value (something I’ve previously covered here). On the cost side, brands that can see and renegotiate their upstream exposure, think agentic AI systems dynamically repricing logistics, materials and freight contracts – will have an advantage far more tangible than any ESG badge.

LVMH’s recent struggles in its wine and spirits division are a reminder that when supply chains misfire or demand shifts regionally, you cannot recycle your way out of a margin squeeze; you have to re‑route inventory, reset terms and re‑plan production in real time. On the revenue side, Digital Product Passports open the door to authenticated resale at scale: if a bag carries a cryptographically secure, DPP‑linked provenance record from day one, its residual value can be higher and more liquid, reinforcing pricing power and desirability over the long term even if it never appears as “sustainability revenue” on a P&L.

So the paradox heading into 2026 is this: sustainability is non‑optional and increasingly codified in law, yet the brands delivering the best financial outcomes are not the ones scoring highest on ESG indices. The real battleground is shifting to supply chain and data: who can build the infrastructure to comply with DPP, strip cost out of complex networks and turn traceability into higher resale value. At the same time, the front end still has to stay focused on what has always moved luxury: product, story and status.

AI as a Means, Not the End

For luxury, the AI play is narrower than in mass fashion. AI will mostly live in back‑end efficiency rather than front‑end marketing, because a high‑touch, human creative layer still defines campaigns and clienteling.

For conglomerates like LVMH, AI is already improving supply‑chain performance via predictive routing, regional vendor‑risk analysis and sharper inventory and demand forecasting. Several houses are also training models on their archives to help designers generate brand‑native ideas, a pattern I flagged in 2023. Online customer service remains a weak point for many luxury brands, and AI agents will not only raise baseline service quality but also enable new concierge‑style experiences that remember preferences and history for VIP clients.

In some ways, luxury brands are not moving fast enough. Third‑party AI personal‑shopper platforms like Daydream (essentially “Pinterest meets Amazon for fashion discovery) risk inserting themselves between maisons and their customers, and are already aggregating behavioural data across thousands of brands at a scale individual houses cannot match. The brand that controls the richest data will have the edge in the age of AI, and while ultra‑high‑net‑worth clients may stay with in‑house advisors, more price‑sensitive aspirational shoppers can be steered toward whichever label is best surfaced, or discounted, by these intermediaries.

AI also adds a new dimension to advertising strategy. On one side is Generative Engine Optimization, where the goal is to be the brand that AI shopping agents and chat‑search surfaces recommend, driven by user sentiment, reviews and material‑quality signals. On the other is physical experience, and as content gets commoditized by AI, live shows, installations and in‑store theatre become even more important differentiators that algorithms can point to but not replace.

Funding flows suggest where AI’s impact will show up first, in infrastructure around luxury rather than in the maisons themselves. Oura, XPANCEO, Eight Sleep, SoftWear, BNTO and Nibertex sit in health data, AR lenses, sleep optimization, robotics, AI‑enabled rental and advanced textiles, and have raised far larger rounds than pure “luxury content” startups, underscoring that the early value is in materials, supply chain and data pipes rather than campaigns. If conglomerates do not build or partner into this stack, these intermediaries will own the customer insight and the AI‑driven margin upside. LVMH’s own 2025 Innovation Award slate points in the same direction: the highlighted startups cluster in AI and data intelligence (Kahoona, OMI), sustainability and circularity (Genesis, Aectual), and supply‑chain and media infrastructure (Authena, Ircam Amplify), not front‑end storytelling tools.

For luxury brands, the end goal with AI is straightforward: expand margins and redeploy the savings into experiences, more distinctive product, better service and richer physical worlds that preserve the category’s scarcity and status.

Where Luxury Goes Next

At the end of the day, luxury is not leather, diamonds or watches. From first principles, luxury is something scarce, something we do not see often. There are two attributes that matter most, desirability and scarcity. A desirable but non‑scarce good like air is not a luxury, and a scarce but non‑desirable good, like an out‑of‑fashion logo‑heavy sweatshirt, is not either. Luxury brands can manufacture desire or manufacture scarcity, but the winners pull on both levers.

The next wave of luxury startups will emerge in the gaps legacy houses cannot or will not fill. One group will build local luxury on Asian cultural codes rather than imported European ones. Another will act as experience facilitators, designing elitist but modern clubs (tennis, racquet sports, wellness) and unique destination real estate such as Lapland‑style retreats for VIPs. A third will be tech partners that own the AI layer, offering hyper‑personal concierge, data‑rich personal shoppers and dynamic drops without fully disintermediating the brand.

On the product side, the big white spaces are health as luxury (genome‑ or biomarker‑based beauty, longevity programs, and performance wearables that shift status from “things” to “self”), quiet‑luxury brands built on drop‑only models to limit inventory risk while supporting secondary prices, and a truly superior non‑animal material that can credibly replace exotics for houses like Hermès. These are the next frontiers where innovation meets aspiration, expanding the definition of what luxury can signal.

The target customer for these categories looks less like an old-world logo collector and more like the kind of consumer backing brands such as Gentle Monster, Miu Miu, or Arc’teryx’s Veilance, seeking edge, scarcity, and cultural relevance alongside heritage rather than logos alone. The double-digit online growth projected for Gentle Monster and Miu Miu’s aggressive expansion suggest there is still plenty of room for new tiers of luxury at different price points, as long as the product feels like the right, resonant one.

Strategic Implications by Brand

Hermès:

Keep playing your own game, but push it one notch further into “life system” rather than “object.” Use the 40%‑plus margins and roughly €12.8 billion net cash to keep investing in ateliers, materials science and your forests and zero‑deforestation agenda, while expanding the world around the product with traveling craft exhibitions, Mystery at the Grooms’ style experiences and quietly elite sport and wellness touchpoints that make Hermès feel like a way of living, not just a thing you buy.

LVMH:

Lean into more personality and less clinical perfection. Use high‑brow culture, ideas and unexpected collaborations to speak to younger clients who want to feel intellectually as well as aesthetically engaged, and extend your existing art‑driven flagships and hospitality ecosystems into new resonant formats, from members‑style tennis and arts clubs to ultra‑rare, story‑rich travel residencies for ultra‑VIPs that make the brand feel like a world they briefly inhabit, not just a logo they wear.

Prada Group:

Treat Prada as the thinking person’s Gucci and push harder into that space, high‑brow fashion with an urban edge, while treating Miu Miu less as a fad and more as a movement by listening obsessively to its core customer. The job is to sustain Miu Miu’s momentum, re‑accelerate Prada and turn around Versace without distraction, using your structural advantages in retail control, factories and disciplined stewardship.

Richemont:

Keep doing what you do best, but make it feel even more like culture and even less like retail. Use the jewelry engine and roughly €6.5 billion net‑cash buffer to double down on museum‑grade moments, Cartier and Van Cleef & Arpels jewelry worlds and Vacheron Constantin at the Louvre style exhibitions as the primary marketing channel, while quietly fixing Specialist Watchmakers’ economics with sharper SKU discipline, fewer mid‑tier points of sale and more true haute horlogerie.

Chanel:

Protect the mystique but modernize the interface. Let creative direction lead with material intelligence and silhouette rather than logos, while using AI quietly to remove friction in beauty and clienteling, Lipscanner style tools and CHANEL & moi grade service, without breaking the white‑glove spell. The goal is to make boutiques feel like cultural events and digital touchpoints feel invisible, so desirability, not distribution, remains the centre of gravity.

Moncler:

Move further away from hype‑beast territory and back to your mountain roots. Luxurious Patagonia is a better north star than the next drop; double down on DTC control, experiential shows and technical innovation so you stay ahead of Arc’teryx type competitors and build a brand that feels refined and durable rather than reactive.

Kering:

Go back to your roots, La Famiglia is a good start, and avoid replaying the 2000s hype cycle. You need your Tom Ford moment, which will not come from following trends but from setting them, backed by serious investment in AI and innovation where peers are already ahead.

Put simply, the recent winners have followed a fairly consistent playbook across different positions in the market.

Over the next cycle, the real divergence will come not from who can raise prices the fastest, but from who can keep compounding relevance while their customer base fragments. The brands that treat scarcity, creative focus and balance sheet strength as operating disciplines rather than marketing lines will still be standing in 2030. The rest will spend the decade relearning how to grow.

For Investors

On today’s numbers, the market is still paying a premium for certainty. Hermès and Richemont are the clearest quality compounders, Prada and Moncler offer the best risk‑reward in listed names, and Kering has effectively become a leveraged turnaround bet on Gucci.

The full valuation tables, risk grid, and 2026–2030 bull, base, and bear scenarios are available in an interactive dashboard for paid subscribers.

Prefer a visual, magazine‑style version of the free essay, with curated charts and brand breakdowns? Read it here: https://luxury-report-2026.theinterneteconomy.xyz/

If you want to work with me on strategy (brand resets, investor theses, AI roadmaps in luxury) or deep dives like this one, message me at jurgis@evoaai.com or on LinkedIn.

Read more

Major banks are no longer confined to innovation labs. They are deploying live pilots with real balance sheets and billions on the line. January 2026 headlines from the World Economic Forum spotlight JPMorgan’s JPM Coin expanding onto public blockchains for seamless 24/7 payments, alongside Citigroup’s Token Services enabling near-real-time cross-border clearing. This shift extends beyond banks. Infrastructure and payments companies like Stripe and Fireblocks are rolling out crypto-native settlement and network layers built for institutional scale.

These are no longer experiments. They reflect a growing reliance on DeFi rails where they outperform legacy systems.

But as capital floods in, one question matters more than any other for builders and investors alike:

Where is this yield actually coming from?

Subscribe now

1. Yield Comes From Lower Operating Costs

It’s no revelation that traditional retail banks offer negligible APYs on consumer deposits.

The core big-bank business model is built around net interest income: borrowing money at effectively zero through deposits and lending it out at meaningfully higher rates. Interest income typically accounts for roughly 70% of bank revenue, with the remaining 30% coming from non-interest income.

That non-interest income shows up everywhere: monthly account fees, overdraft fees, ATM and wire fees, debit and credit card interchange, credit card interest, and commissions on insurance and investment products.

Borrowing at 0% and lending at 8%+ is a very good business.

Why are fintechs able to offer higher rates?

The primary reason is lower operating costs. Fintechs like Revolut, now a ~$75B company, operate with no physical branches and far leaner cost structures. That efficiency allows them to pass through a larger share of interest income, often earned on Treasuries, directly to customers. It’s their core customer acquisition strategy, as yield becomes a marketing expense designed to pull deposits away from banks.

Fintechs also reprice faster. When the Federal Reserve hikes benchmark rates, fintechs tend to adjust savings rates quickly. Large banks, constrained by internal bureaucracy and legacy systems, move slowly. Over time, that lag compounds.

Many fintechs are also willing to lend deposits into higher-yielding, riskier asset classes than traditional banks, further expanding the rate spread they can offer to users.

As a result, interest-bearing deposits have become table stakes across fintech and neobank platforms, and depositors are increasingly demanding utility from idle balances rather than accepting zero yield.

It’s worth noting that most fintechs still don’t hold deposits themselves. Without a banking license, they’re required to place customer funds with licensed partner banks. Yet even within that constraint, fintechs are often able to negotiate higher APYs for their users through a combination of scale and economics:

• Bulk deposit negotiation

• Revenue-sharing agreements

• Lower servicing costs

• Higher interchange capture on debit and credit cards

• Cross-sell incentives

This is where stablecoins push this model to its logical extreme.

Stablecoins resemble fintechs on steroids: lean teams, minimal operating expenses, and a far greater ability to pass through Treasury yields directly to users.

As yield distribution (directly or via loopholes) gets normalized via stablecoin as a service providers like M0, Agora, Paxos and Bridge, capturing the customer mindshare will require

a) existing distribution, or

b) growing distribution via marketing (which increases operating expenses).

Just like the Ethereum blockchain is optimized for decentralization, big banks are optimized for capturing net interest margins and maturity transformation.

It’s ultimately a design choice.

2. Yield Comes From Lending to Traders

Yield from passing on Treasury rates is the most vanilla form of DeFi yield. It works, but it has a ceiling. Over time, those rates compress, and the advantage fades.

DeFi’s more interesting yield sources begin once capital is put to work inside markets rather than parked against government debt.

One of the most important drivers of DeFi yield is trading activity.

Across lending markets, stablecoin borrowing rates tend to rise alongside ETH prices. As prices move higher, demand for borrowing stablecoins increases, pushing lending APRs up in tandem. The relationship is not perfect, but the directionality is clear.

Several forces reinforce this dynamic:

• Rising asset prices increase collateral value, which expands borrowing capacity and enables larger positions. Borrowed stablecoins are often redeployed into additional market exposure, reinforcing price momentum.

• The opportunity cost of lending stablecoins increases during strong market moves. Holding ETH directly becomes more attractive, so lenders require higher compensation to remain in stablecoins.

• Increased demand for leverage. Stablecoins serve as the primary margin asset across perpetual futures markets, supporting both long and short positioning. People longing start to think it’s the beginning of a super cycle, people shorting think it can’t go up that much more

A growing share of stablecoin yield is also tied to liquidity provision strategies that depend directly on trading volume (among other things). On Jupiter, 75% of liquidity provider fees come from opening and closing positions, price impact, borrowing costs, and trading fees. As leverage and activity increase, these fee streams expand. When volumes fall, yields compress

3. Yield Comes From Delta-Neutral Trading

Another major source of DeFi yield comes from delta-neutral trading strategies.

The popular example is Ethena’s sUSDe, a synthetic stablecoin backed by a delta-neutral structure rather than cash reserves. The system holds collateral such as ETH or stETH while simultaneously opening short positions in derivatives markets. The short offsets directional price exposure, while funding rates from perpetual futures generate income. The USD peg is maintained through continuous hedging rather than redemption against fiat.

A similar dynamic exists in RLP, the leveraged yield-bearing token developed by Resolv Labs. RLP functions as a risk buffer for the USR stablecoin, absorbing market and counterparty losses that arise from delta-neutral hedging. In exchange, RLP holders capture higher returns sourced from staking rewards and funding fees.

In both cases, yield is produced by managing basis risk and market structure rather than by lending against fixed collateral. The returns compensate holders for exposure to hedge execution risk, funding rate volatility, and counterparty dynamics across derivatives venues.

This form of yield sits structurally between lending and active trading, offering higher returns at the cost of greater operational and market sensitivity.

4. Yield Comes From the Wealth Effect and Tax Avoidance

A useful example here is Gauntlet’s largest vault, USDC Prime on Base. The vault allocates nearly all of its assets to the cbBTC/USDC pool, lending out USDC against cbBTC collateral at roughly 4%.

The obvious question is who sits on the other side of that trade. Who is borrowing USDC against Bitcoin collateral? While some of this borrowing could be attributed to desire for leverage, because cbBTC is Coinbase’s wrapped Bitcoin product, it could be inferred that a significant portion of borrowers of USDC do so for tax purposes, while preserving investment in Bitcoin.

Selling Bitcoin triggers capital gains tax, while borrowing against it doesn’t. This is a familiar strategy in traditional wealth management, where high-net-worth individuals routinely borrow against equities, real estate, or other appreciating assets rather than selling them.

As Bitcoin prices rise, this dynamic strengthens. Higher prices increase collateral value, which expands borrowing capacity and pushes utilization higher.

In this case, yield is driven less by speculative trading and more by balance-sheet behavior tied to asset appreciation. The wealth effect increases borrowing power, and borrowing demand sustains yield even in relatively calm market conditions.

5. Yield Comes From Looping

Looping is one of the most direct ways DeFi turns leverage into yield. At a basic level, it involves repeatedly borrowing against collateral and redeploying the borrowed capital, constrained by loan-to-value (LTV) limits.

In its simplest form, looping uses volatile assets like ETH as collateral. With an 80% LTV and a borrowing rate around 2.47% APY, a maximally levered loop looks roughly like this:

• Loop 1: Deposit $1,000 of ETH → borrow $800 in stablecoins → buy more ETH

• Loop 2: Deposit $800 of ETH → borrow $640 in stablecoins → buy more ETH

• Loop 3: Deposit $640 of ETH → repeat

As you can see in the math below, each loop increases effective leverage, with the final leverage ratio determined by the LTV cap and the number of iterations.

This strategy assumes near-maximal leverage and minimal safety margins. In practice, few users run this to the limit. Even modest price drawdowns can trigger liquidations once collateral buffers thin out.

A more structurally interesting, and arguably safer, version of looping uses principal tokens (PTs) rather than volatile collateral.

This approach was enabled by Pendle, which introduced a way to separate yield-bearing assets into principal (PT) and yield (YT) components. PT holders receive fixed yield, while YT holders gain exposure to variable future yield. The market clears because one side prefers certainty and the other prefers convexity.

Instead of posting ETH or BTC as collateral, users can now post fixed-yield stablecoin instruments (such as PT-sUSDe) to borrow. Because these assets exhibit lower volatility, lending markets typically allow higher LTVs, often around 90%.

A maximally levered PT looping strategy looks like this:

• Loop 1: Deposit $1,000 of PT-sUSDe → borrow $900 USDe → buy more PT-sUSDe

• Loop 2: Deposit $900 of PT-sUSDe → borrow $810 USDe → buy more PT-sUSDe

• Loop 3: Deposit $810 of PT-sUSDe → repeat

The effect on yield is material. At maximum leverage, looping PT-sUSDe can increase effective yield by more than 9× relative to holding sUSDe without leverage.

Execution complexity used to be a limiting factor. That friction has largely disappeared. Platforms such as Kamino now offer one-click looping, lowering the barrier to entry and increasing system-wide leverage.

6. Yield Comes From Private Credit

Another growing source of DeFi yield comes from access to private credit markets that were historically difficult, if not impossible, to reach on-chain.

The largest on-chain credit operation today belongs to Figure Technologies, which operates on the Provenance Blockchain. Figure has tokenized assets such as HELOCs and other consumer loans, reaching close to $19 billion in total value originated, with roughly $15 billion active as of early 2026. The yield here comes from traditional private lending economics (credit spreads, borrower demand, and structured underwriting) now made accessible through tokenized rails.

On the institutional side, ACRED has emerged as one of the more visible on-chain private credit feeder funds. ACRED represents tokenized exposure to Apollo Global Management’s Diversified Credit Fund, launched in partnership with Securitize. The underlying fund targets mid-single to low-double-digit returns from private credit. When used as collateral in DeFi lending markets, that yield can be further amplified through looping, pushing effective returns beyond what the underlying credit assets generate on their own.

Some private credit strategies push the abstraction even further. USD.AI presents itself as a stablecoin, but functions more like a credit vehicle. It deploys stablecoin reserves into loans backed by tokenized NVIDIA GPUs housed in data centers, tying yield to demand for AI compute infrastructure rather than financial markets. The structure introduces operational and counterparty risk, but it also opens a yield source that has little direct correlation to crypto trading activity.

Across these examples, the common thread is illiquidity and access. Private credit yields exist because capital is locked up, underwriting is non-trivial, and risks are less transparent than in public markets.

7. Yield Comes From Incentives

A meaningful portion of DeFi yield comes from explicit incentive programs rather than organic market demand.

Yield enters the DeFi system through incentives originating at the protocol and stablecoin levels. The distinction between foundational yield and subsidized yield matters, because much of what appears downstream is built on incentives paid upstream.

At the base of the stack sit protocol-level rewards. Validator rewards on networks like Ethereum and Solana, alongside liquidity provider rewards on DEXes such as Uniswap, Hyperliquid, and Jupiter, form the foundational layer of yield in DeFi. This yield is paid through inflation, priority fees, and MEV, compensating validators and liquidity providers for securing networks and markets.

Liquid staking tokens such as stETH, now the cornerstone of the DeFi ecosystem, turn these protocol-level incentives into composable financial primitives. Once tokenized, staking yield becomes collateral, margin, and a reusable building block across lending, trading, and structured products.

From there, incentives are recycled downstream.

These same protocol incentives, particularly from newer protocols, are often used to subsidize stablecoin and curator yields. A clear example is Gauntlet’s USDC Frontier vault, where roughly 0.32% of the yield comes directly from Morpho incentives layered on top of base lending returns.

USDC Frontier then allocates capital toward tokens advertising higher risk-adjusted yields, including stcUSD, PT-stcUSD, RLP, and permissioned credit-backed assets such as AA_FalconXUSDC. Borrowers frequently loop these tokens to increase exposure, while Gauntlet supplies USDC under the assumption that these instruments behave as stable collateral.

For several of these assets, yield ultimately traces back to incentive programs rather than underlying economic activity:

• stcUSD derives yield from Cap Money’s points program.

• PT-stcUSD converts that incentive-driven yield into fixed returns via Pendle.

• siUSD operates as a stablecoin–curator hybrid distributing rewards to stakers.

Other assets bundled into these strategies draw yield from mechanisms discussed earlier:

• RLP sources yield from delta-neutral trading strategies.

• PT-srUSDe relies on delta-neutral positioning.

• AA_FalconXUSDC earns yield through private credit exposure.

• syrupUSDC ties returns to on-chain private credit markets.

Incentive-driven yield plays a critical role in bootstrapping liquidity and directing capital. It accelerates adoption, but it behaves differently from lending or trading-based yield. Its durability depends on continued emissions, point programs, or external subsidies. When those incentives fade, capital reallocates quickly.

8. Yield Comes From Pseudo “Hedge Funds”

Not all stablecoins are built to behave like money. Some function as wrappers around trading strategies, effectively tokenized shares in relatively stable hedge-fund-like vehicles.

There are more reputable implementations, such as Ethena, but the category as a whole carries real risk. The collapse of Stream Finance made that explicit, and even Ethena has experienced material drawdowns during periods of market stress.

Looking at the highest-yielding stablecoins, a clear pattern emerges. Names like Avant, Unitas, and Neutrl consistently sit near the top of the yield rankings.

• Unitas’s sUSDu generates yield through delta-neutral JLP positions, hedging long exposure with short perpetuals.

• Avant’s savUSD earns yield through actively managed, market-neutral strategies including basis trades and lending rate arbitrage.

• Neutrl’s sNUSD functions as a tokenized hedge-fund-like strategy, capturing alpha by purchasing locked or OTC altcoins at a discount and hedging price exposure with perpetual futures.

Across these products, yields have been volatile and, over time, compressing. Early high returns attract capital, strategies crowd, and marginal returns decline.

In effect, these resemble a more structured evolution of DeFi Summer 2020 yield farming. The strategies are more sophisticated, but the underlying dynamics are familiar. They offer access to yield sources that traditional finance struggles to reach, while introducing distinct risks:

• Imperfect hedging, including exposure to auto-deleveraging mechanisms on derivatives venues

• Strategy execution failures, including periods where hedges are incomplete or misaligned (e.g. Stream Finance)

• Temporary APYs driven by mercenary capital rather than durable demand

Protocols have responded by improving transparency through dashboards detailing asset composition, counterparties, and strategy mechanics. That helps, but it doesn’t eliminate structural risk.

Two pressures are harder to escape. Successful strategies crowd quickly, forcing teams to search for new sources of alpha. At the same time, exchanges increasingly internalize these strategies and offer them directly to users at lower cost. Jupiter’s native delta-neutral JLP product is a clear example. Many high-yield stablecoins rely on the same mechanics without the same distribution or fee advantages.

Calling these products “stablecoins” is debatable, however, packaging fund exposure into a stable wrapper gives users a cleaner unit of risk. For example, if an underlying strategy, expressed through a stablecoin, yields 8%, a holder seeking higher returns (and willing to accept more risk) can loop that position (see Section 5). In some cases, this produces alpha: leveraging a lower-volatility, lower-yield strategy can deliver better risk-adjusted returns than simply allocating to a higher-APY product outright.

What’s being offered here is not free yield, but balance-sheet engineering, increasingly accessible to anyone willing to understand the machinery underneath.

What Can Go Right and What Can Go Wrong

By now, the mechanics are clear. The remaining question is what conditions actually push yields higher and where the system may break.

How Do We Get Higher Rates?

• Higher treasury rates → higher base yields passed on to stablecoin holders

• Higher trading volumes → higher demand for leverage → higher borrowing costs → higher yields for lenders

• More traders losing money → higher profits for counterparties to leveraged perpetuals → higher JLP (Jupiter Liquidity Provider) and HLP (Hyperliquidity Provider) returns → higher APYs for stablecoins sourcing yield from these strategies

• Stronger wealth effect (or anticipation of higher BTC prices) → higher unrealized gains → greater borrowing capacity against Bitcoin and other assets → higher cbBTC-backed lending yields

• Looser risk parameters → higher loan-to-value ratios → more looping → higher effective yields

• Access to private markets (often less liquid and riskier) → illiquidity premium → higher yields

• Larger crypto protocol airdrops → higher incentives → increased demand for Pendle YT → higher fixed rates on Pendle PT

• 

Where This Could Go Wrong

• Depeggings, even temporary ones, can force looped positions to unwind. The drawdown experienced by Ethena after October 10th, 2025 is a clear example.

• Private credit–backed stablecoins face liquidity mismatches (e.g. 31-day redemption windows) and counterparty risk that is difficult to observe on-chain.

• Lower trading volume compresses funding rates, reduces leverage demand, and directly lowers yields tied to trading activity.

• Opacity in vault demand makes it difficult to assess risk. In some cases, it’s unclear what share of USDC borrowing comes from independent market participants versus the protocol whose asset is used as collateral (e.g. how much of outstanding stcUSD borrowing originates from the Cap Money ecosystem itself).

• Stablecoins as funds in disguise. Many of these structures resemble curators, trading funds, or private credit allocators where the stablecoin represents a claim on a balance sheet. Like earlier generations of crypto allocators, they can fail, triggering losses that ripple back to token holders and connected protocols, as seen with Stream Finance.

• Yield competition creates reflexive risk-taking. Protocols that compete primarily on yield become dependent on maintaining that yield to defend TVL. When organic returns fall, the incentive is to move further down the risk curve. Ethena serves as a cautionary example.

Even with unresolved transparency issues, DeFi has introduced financial instruments that, when used correctly, can offer higher risk-adjusted returns than traditional markets. It’s no surprise that Wall Street is eager to gain access.

Conclusion

Most DeFi-native yield still traces back to trading activity.

Whether it shows up as lending to traders, looping, delta-neutral strategies, or pseudo hedge-fund stablecoins, trading volume and leverage remain the primary engines of yield. Even strategies that appear market-neutral ultimately depend on active markets, funding rates, and counterparties taking the other side of risk.

Beyond trading, the remaining drivers of yield fall into three buckets.

1. Macro conditions set the floor. Treasury pass-through and protocol staking rewards anchor the low-risk end of the spectrum, but on their own they produce modest returns.

2. Speculation and incentives amplify yield temporarily. Emissions, points programs, and subsidized vaults accelerate adoption and liquidity, but they are inherently transient.

3. Illiquidity and balance-sheet risk push yields higher. Private credit, structured products, and hedge-fund-like strategies compensate users for locking capital, absorbing volatility, or warehousing tail risk.

What changes outcomes is not the source of yield, but how aggressively it is multiplied.

Looping, PT structures, leveraged incentives, and multi-strategy vehicles all increase yield per unit of capital by increasing risk per unit of capital. The charts make this clear: as strategies move up the risk curve, returns rise not because the underlying yield source improves, but because leverage, complexity, and fragility increase.

This is why lower-risk strategies like plain lending, treasury pass-through, and unlevered staking, consistently offer lower returns when used in isolation.

The core insight is this:

DeFi does not create yield out of thin air. It rearranges, amplifies, and repackages risk.

Understanding where yield originates, and how it is multiplied, is the difference between owning an instrument and understanding the machine behind it.

If you’re thinking about these questions, let’s talk.

LinkedIn: https://www.linkedin.com/in/jurgis-pocius/

Email: jurgis@evoaai.com

Join for emerging tech insights from strategy to implementation. Trusted by JPMorgan, McKinsey, HBS, Dior, Nike and leading VC funds and traders. Because after all, Jeff Bezos found the idea for Amazon via a newsletter just like this.

Subscribe now

We all hear that AI is “changing the world” but most of the advice for investors is surface-level stuff.

I wanted something practical. I’ve launched Automated Alpha to bridge that gap.

I’ll keep writing here, but Automated Alpha is where I’ll share practical AI workflows for VCs, professional investors, retail and crypto traders.

The first deep dive, how I built an automated stock analyst that generates 14-page investment memos in under 5 minutes.

The output is amazing and will only continue to improve.

Apple Investment Review Outlook

15.7MB ∙ PDF file

Download

Download

This Apple report was generated entirely by the tool. If you want the prompt workflows and the tech stack behind it, you can find the full breakdown below.

Automated Alpha

How I Built an Automated Stock Analyst (And How You Can Too)

If you’re interested in how I was able to generate all three in under 5 minutes, then keep reading. The presentation and notes are especially rigorous in their insights and data…

Read more

3 months ago · Jurgis Pocius

This isn’t a dealflow problem. Most VCs have dealflow. The best funds are seeing 700+ decks a quarter. The issue is triage throughput.

Here’s what actually happens:

Deck hits inbox → someone says “we’ll review this” → founders send a follow-up two weeks later → by the time anyone actually reads it, the round is closed or the founder has moved on.

The best opportunities don’t wait for you to find them. They are often buried under the noise of a thousand average decks, or they move so fast they're gone before you even open the email. Finding a great company is quite literally like finding a needle in a haystack.

If you run or work at a fund that sees 200+ decks a quarter but only has bandwidth to seriously evaluate 20-30, this piece is for you.

I’ve been in crypto and investing since 2017, where I first began navigating high-volume, data-heavy markets. I’ve sat on both sides of the table, evaluating opportunities and watching how funds evaluate them. For the past year, I’ve applied that experience to building AI automation systems for operators.

The VC triage problem kept surfacing in conversations, so I built a system specifically to solve it.

Subscribe now

The Shift Is Already Here

The data is overwhelming. According to Allvue’s 2025 GP Outlook Survey, AI adoption among private market firms hit 82% by year-end 2024, up from just 47% the previous year. The number of data-driven VC firms jumped 20% from 2023 to 2024 alone.

Deloitte’s 2025 M&A Trends Survey found that 97% of respondents have begun incorporating generative AI or advanced data analytics into their dealmaking processes, with digitization rising significantly in target identification (80%) and target screening (79%).

The elite funds have been throwing resources at this problem:

SignalFire raised additional capital in April 2025, significantly expanding its assets under management. Its Beacon AI platform analyzes vast amounts of organizational and workforce data, serving as what founder Chris Farmer described to Crunchbase as “the fabric that stitches the entire firm together.”

EQT’s Motherbrain platform has been instrumental in sourcing several direct investments. It integrates numerous external data sources with the firm’s internal network insights. At their 2024 Capital Markets Day, EQT outlined its goal to “build the most AI-literate investment organization in the world.”

Nik Storonsky, founder of Revolut, co-founded the AI-driven VC firm QuantumLight, which uses proprietary machine learning algorithms to source investments and claims to “almost eliminate human judgment” from the process.

However, some of these systems were built on pre-generative AI architecture. They required massive teams and eight-figure budgets to build. They’re optimized for their specific workflows, not customizable. And whether they’ve actually “solved” triage, or just thrown enterprise resources at it, is an open question.

I also think “human VCs” can’t be entirely automated as relationships and human connection plays a big role in the industry. Instead these automations should be applied to the highest leverage and AI-fit parts.

What’s clear is that the problem is real, the market is moving, and generative AI has changed what’s possible.

The Mixed Results of Early AI

The market is moving. But most implementations reveal how early we still are.

Tim Draper launched an AI “digital twin” that lets founders pitch via text or voice chat. The direction is right, AI handling the top of funnel so humans focus on high-signal deals. But Draper himself admits it’s “flaky.” More fundamentally, the system doesn’t accept pitch deck uploads, missing the structured information that decks provide.

Alliance DAO is the gold standard for crypto accelerators, vetting over 3,000 startups a year. At that volume, they have no choice but to be systematic. However, their entry point relies on a heavy initial application consisting of 39 different questions.

While this extracts deep data, it creates a massive adverse selection risk. The “hottest” founders, the ones with multiple term sheets and zero time to spare, often won’t fill out 40-question forms. They don’t have to. This means the most systematic funds can end up seeing the most desperate founders, while the elite talent bypasses the “inbox” entirely.

The Case for a Zero-Touch Interview

What’s missing across the board is a Zero-Touch Interview layer.

Instead of forcing the founder to do the manual labor of data entry, a modern system should be able to perform a Key Claims Audit autonomously. By extracting data from the deck, verifying team pedigree via LinkedIn, and cross-references technical claims against public data, a fund could get the same depth as a 15-minute introductory call, without either party ever picking up the phone. You get high-level signal with zero-click friction.

The pattern is clear: right problem, incomplete solutions. Most funds are stuck choosing between a “flaky” chatbot or a form that scares away the best talent.

Historically, “quant” funds like SignalFire or EQT had to spend millions building custom models just to understand unstructured text. Today, Generative AI allows any fund to quantify the qualitative, turning messy pitch decks into structured, comparable data without a dedicated team of data scientists.

You no longer need an eight-figure engineering budget to have a quantitative edge; you just need the right architecture.

I spent the last month architecting the solution myself: The 60-Second Triage Engine.

The 60-Second Triage Engine

An end-to-end workflow that turns inbound startup decks into scored CRM entries with thesis-aligned analysis. The whole thing runs in sub-60 seconds per deck 24/7.

What the Startup Sees

A simple form (I used Typeform, any other form e.g. Tally works as well) with as few or as many questions as you like. Crucially, it includes an area to submit the pitch deck.

What the VC Sees

All of this occurs without any human input beyond the startup’s initial submission:

1. Automated Research Layer

The system extracts text from the PDF, pulls LinkedIn data, scrapes the company website, and cross-references claims against public data sources. It’s doing in seconds what would take an analyst 15-20 minutes.

It then uploads this information directly to your CRM, whether that’s HubSpot (as shown in the images below) or a simple Google Sheet, so you can easily track all submissions and revisit them when needed.

To show you what this looks like in practice, I ran a publicly available deck through the system, RoboForce, a robotics company that raised $10M in 2025:

2. Investment Memo Generation

This gets passed to the CRM as seen above, but it didn’t fit in the screenshot.

RoboForce — Investment Memo (January 22, 2026)

Conviction Level: Med

The Thesis (1 sentence):
RoboForce is a deep-tech robotics company leveraging proprietary physical AI (RF-Net) and an elite team to automate labor in harsh industrial environments, but the investment risk hinges entirely on resolving the conflicting claims regarding current commercial deployment status.

Key Claims Audit:

Claim: RoboForce is developing autonomous, mobile manipulation robots (“Robo-Labor”) named TITAN.
Status: SUPPORTED
Sources: Pitch Deck; Website

Claim: The core IP is a unique “AI Expert Model with 1mm Accuracy” for complex manipulation tasks (Pick, Place, Press, Twist, Connect).
Status: SUPPORTED
Sources: Pitch Deck; Website

Claim: Founder Leo Ma is a high-caliber executive with a track record of taking a company (CYNGN) to Nasdaq.
Status: SUPPORTED
Sources: LinkedIn Analysis

Claim: The team includes engineers and leaders from Tesla, Amazon Robotics, Google, Waymo, and CMU/UMich.
Status: SUPPORTED
Sources: Pitch Deck; Website

Claim: The company is planning or executing an “Alpha Robot Onsite Test” in 2024, indicating a pilot stage.
Status: INCONSISTENT
Sources: Pitch Deck (contradicts Website)

Claim: The company is “Production-ready with active commercial deployments” and initial pilots are already active.
Status: INCONSISTENT
Sources: Website (contradicts Pitch Deck)

Claim: RoboForce has secured more than 11,000 robot orders through letters of intent (LOI).
Status: SUPPORTED
Sources: Website

Claim: Institutional backing includes Myron Scholes, Gary Rieschel, and Carnegie Mellon University.
Status: SUPPORTED
Sources: Pitch Deck; Website

Claim: Total funding is $15 million, including a $10 million early-stage announcement.
Status: SUPPORTED
Sources: Website

Claim: Current ARR, Burn Rate, and Gross Margin metrics are not disclosed.
Status: SUPPORTED
Sources: Pitch Deck


Bull Case:
- Exceptional Pedigree and Founder-Market Fit: The founding team has direct experience building and scaling foundational AI/Robotics technology (Baidu, CMU) and successfully taking a company public (CYNGN), lending strong credibility to the technical claims.
- Differentiated Deep Tech IP: The company is developing a full-stack, proprietary solution including the TITAN hardware and the RF-Net foundation model, claiming 1mm precision and continuous learning capability, positioning it as a foundational platform for "Physical AI."
- Strong Indicated Market Demand: The company explicitly targets a massive and worsening labor shortage in utility-scale solar construction and has secured over 11,000 robot orders via LOI.
- Authoritative Validation: The product has been highlighted by NVIDIA CEO Jensen Huang at GTC, and the founder has been recognized as a WEF Technology Pioneer (Future recognition).

Bear Case:
- Contradiction on Commercial Status: The most critical risk is the INCONSISTENCY between the Pitch Deck (2024 Alpha Test/PoC stage) and the Website (Production-ready/Active commercial deployment). This confusion suggests a misalignment on execution timeline or marketing vs. reality.
- Unverified Financial Health: Core financial metrics required for evaluating a capital-intensive deep-tech play (ARR, Burn Rate, Gross Margin, Churn) are Not Disclosed.
- Execution Risk: This is a full-stack, proprietary hardware and software venture, demanding significant, sustained capital and flawless execution across complex mechanical engineering, AI, and control systems.
- Lack of Commercial Proof: While 11,000 LOIs are cited, the company has not provided verifiable customer logos, signed contracts, or revenue figures to substantiate the claimed active pilots.

Power Law Test: PLAUSIBLE

Hard Truths (3 questions):
- Which statement regarding product maturity is accurate: the Pitch Deck's "Alpha Robot Onsite Test 2024" or the Website's "Production-ready with active commercial deployments"?
- What is the current burn rate and capital runway, given the $15 million raised and the evident high R&D cost of full-stack hardware/software development?
- Can the company immediately provide signed customer contracts or verifiable revenue figures to substantiate the 11,000 robot letters of intent (LOI) and active pilot claims?

3. Thesis-Aligned Scoring

The model is trained using two mechanisms:

First, a prompt/context window where you define your investment criteria directly. Your sector focus. Your green and red flags. You assign weights based on what actually matters to your fund.

For instance, this was used for RoboForce, but this could be made much more comprehensive:

## Weights
1) Team–Market Fit (40%)
2) Traction vs. Claims (30%)
3) Verification Score (30%)

## Tier Thresholds
- Priority: score >= 75
- Standard: score 50–74
- Pass: score < 50


# Here are some examples of startup scores that I have made previously:

{{ JSON.stringify($json.examples, null, 2) }}

Second, a vector database trained on your portfolio companies and the scores you’d retroactively assign them. Did successful and unsuccessful investments have common themes? What was the industry? Was the founder first-time or not? This “trains” the system on pattern recognition specific to your fund’s actual history.

In practice, the vector database performs a semantic search across your entire deal history, surfacing 'lookalike' companies to see how this new opportunity aligns with your previous investment decisions.

Below is a view of the sample vector database used for our run.

4. Automated Routing and Slack Alerts

Every deck gets a score (0-100) and a tier:

• Tier 1 (75+): Slack alert fires immediately. Strong fit, worth deep dive.

• Tier 2 (50-74): Queued for review. Possible fit, review when bandwidth allows.

• Tier 3 (<50): Auto-archived with rationale. Outside thesis.

VCs receive a Slack notification, while top-tier startup applicants get instant emails or SMS messages with a Calendly, Zoom, or Google Meet link. Lower-scoring applicants instead receive a personalized thank-you email.

Why This Changes the Game

Here’s what shifts when this system is running:

1. Response time drops from weeks to hours

Founders get a reply the same day they submit. Even if it’s “not a fit right now,” they know and may even get automatic feedback. This alone changes how founders perceive your fund.

2. Partners stop drowning in noise

If you’re only seeing Tier 1 alerts, you’re not wading through 200 decks to find the 3 that matter. You’re seeing the 3 and can go deep immediately.

3. You stop losing deals to lag

The “we’ll circle back” problem goes away. If it’s Tier 1, you know within an hour. If the founder is raising on a tight timeline, you’re not finding out two weeks later when the round is closed.

4. Pattern recognition improves

Because everything is logged consistently, you start seeing patterns. Which sources send the best deals (e.g. if you have multiple forms). Which sectors are overrepresented in your inbound. Where your thesis is too narrow or too broad.

5. Junior team members become more effective

Instead of “read these 30 decks and tell me which are interesting,” it’s “here are the 5 Tier 1s from this week, do deep research on these specific questions.”

The funds implementing AI for triage will compound their advantages: faster response times → better deal access → better returns → better brand → more inbound → better LP terms.

You don’t need eight figures and a dedicated AI team to build this anymore. Generative AI changed the equation. You need the right architecture and someone who knows how to configure it for your thesis.

Breaking Down the Numbers

Here's what the numbers look like for a mid-sized fund processing 2,000 decks per year (roughly 500 per quarter):

Direct Cost Savings

• Decks/year: 2,000

• Total triage cycle saved: 10 minutes/deck (Skim + Research + CRM Entry)

• Blended hourly cost (90% Associate @ $156/hr, 10% Principal @ $233/hr): $164/hr

  ~$55K/year in direct labor savings

Opportunity Cost Unlocked

• ~333 senior hours freed annually

• Additional high-quality founder interactions enabled: ~166 per year

• Conservative conversion to investment: ~2%

• Incremental investments enabled: ~3 per year

• Expected value per investment: $250K–$500K

  Implied expected opportunity value: ~$0.8M–$1.6M

The labor cost is intuitive, but the opportunity cost is where it gets brutal.

Where This Goes Next

This architecture is a foundation, not a finish line. The same logic of quantifying the qualitative can be extended across the entire lifecycle of a fund:

• Async Portfolio Coaching: Startups can submit pitch decks or strategy docs and receive instant, thesis-aligned feedback via Slack or email. This allows a fund to deliver value at 2:00 AM with zero marginal cost.

• Partner Proxies: Imagine an AI node trained on a specific partner’s writings, investment memos, and recorded conversations. Each partner has their own proxy node that discuss investment decisions while the real partners are asleep.

• Systematized Brand Building: You can turn partner meeting notes into high-quality LinkedIn posts and blog content automatically. In an era where “VC as a Media Company” is the winning strategy, the funds that systematize their content pipeline will win the attention game.

• Institutional Trading Journals: For hedge funds or liquid token funds, this system can maintain a fund-wide journal of entry conditions, reasoning, and outcomes. You can query the vector database before entering new positions to see where similar logic failed or succeeded in the past.

• Founder Intelligence Audits: By analyzing transcript data from initial calls (via tools like Otter, Gong or Grain), the system can extract “unwritten” data: founder personality traits, potential blind spots, and communication styles. This provides a level of psychological depth that traditional quant models simply cannot index.

The goal of these extensions is simple: to extract the success patterns hidden in your fund’s unstructured data.

Who This Is For

• You run or work at a fund that sees 200+ decks a quarter but only has bandwidth to seriously evaluate 20-30.

• You’re tired of founders going cold because their deck sat in someone’s inbox for three weeks.

• You want dealflow organized by fit, not by which associate happened to check email first.

• You recognize that the current system, serial human processing of every inbound deck, doesn’t scale and is leaving money on the table.

The goal isn’t to eliminate human judgment. It’s to make sure human judgment is spent on deals where it actually matters.

If you want more content like this, I created a new Substack for tactical tips fort investment professionals. I want to keep The Internet Economy more strategic, while Automated Alpha more tactical.

How I Partner with Funds

I don’t believe in one-size-fits-all SaaS for venture capital. Every fund has a unique thesis, specific green flags, and a distinct culture. A generic tool cannot capture the nuance of your investment committee.

Instead, I work with a small number of funds to build proprietary intelligence layers that they own forever. My work falls into two categories:

• Custom Implementation: I architect and ship an end-to-end triage engine (like the one described above) tailored to your specific stack and CRM. This is a fixed-scope engagement that typically takes 2 to 3 weeks.

• Strategic AI Advisory: For funds looking to rethink their entire digital infrastructure, I provide ongoing consulting. We look at everything from automated sourcing to portfolio monitoring and LP reporting.

I prioritize quality over volume. I only engage with 1 or 2 new funds each month to ensure the systems I build are robust and truly aligned with the GP’s vision.

If you are ready to modernize your triage process or want to discuss a broader AI roadmap for your firm, let’s talk.

You can reach me directly at jurgis@evoaai.com or DM me on LinkedIn.

Join for emerging tech insights from strategy to implementation. Trusted by VCs, LVMH, JPMorgan, McKinsey, Nike and more. Because after all, Jeff Bezos found the idea for Amazon via a newsletter just like this.

Subscribe now

Memes are now tradable assets with eight-figure market caps. Prediction markets surged to $13-15 billion in annual trading volume by late 2025, up from just $100 million in early 2024. “Vibe coding” platforms like Replit are selling the dream that anyone can build the next Silicon Valley unicorn through natural language prompts. Your college roommate is selling a testosterone optimization course from a rented Lamborghini. The kid who barely passed economics is running a peptide protocol coaching empire promising fitness results “without the effort.”

This isn’t just “late-stage capitalism” social media discourse. Something deeper is happening. The question that keeps me up at night isn’t whether everything is becoming financialized the data confirms it already has. The question is: what does this tell us about where we are as a culture, and what should we actually do about it?

Subscribe now

Three Forces that Will Define 2026 and Beyond

1) “Secure the Bag” Before AGI

Some think we’re 2-5 years away from AGI, others think it’ll take decades. Whether or not this timeline is accurate matters less than the behavioral shifts it’s producing.

I’m seeing incredibly talented people abandon skill development for “bag securing.” The logic is simple, if brutal: why spend five years mastering a craft when AI might do it better in three? Better to extract maximum value from the current system before it fundamentally restructures. In some ways, this is very logical.

You can see this urgency playing out everywhere. The meme coin market crashed from peaks around $87 billion in trading volume, only to be replaced by prediction markets like Kalshi and Polymarket. Same mechanics, different wrapper. Insider groups profiting early, retail making money initially, then losing it all. The cycle is repeating.

What people often don’t realize is that the biggest earners aren’t necessarily traders, but affiliate code peddlers. The memecoin insider groups of yesterday are now repackaged as prediction market Discord servers and betting tip groups.

How are these groups able to attract new participants if at least 70% are losing money? The answer is aggressive affiliate marketing creating the illusion of easy wins. These sports betting and gambling groups are the most prolific affiliate commission payers, often giving out 30% commission rates.

While I’m not opposed to coaching or groups in general, as long as there’s real value being provided, this often feels like a house of cards.

The AGI deadline creates a weird psychological state. It’s like knowing the party ends at midnight. Everyone optimizes for short-term extraction rather than long-term relationship building. And you can feel this in how people approach work, relationships, and time itself.

Having navigated the memecoin wave and the broader internet money movements, I’ve realized it was never about the “tech” or betting on culture that made memecoins popular (as many VCs and crypto experts thought and may still think). It was the rush of the underlying “make money online” cohort.

The same group migrated through:

• eBay reselling (2005–2010)

• Affiliate marketing (2010–2014)

• Dropshipping (2015–2018)

• Social Media Marketing Agencies (2017–2020)

• TikTok Shop (2023–2025)

• And now some SMMA operators are (unfortunately) trying to co-opt AI automation (2025→)

This financial independence through internet money has been present for a while. But AGI and AI discussions, paired with our next topic of discussion, have created an atmosphere where internet capital markets feel like the only viable path to wealth.

2) A Regular Job is Now a Path to the New Lower Class

Here’s a number that should terrify you: over the past five years, home price growth has nearly doubled wage growth.

If you’re a teacher, nurse, or skilled tradesperson earning $75,000 (solidly middle class by traditional definitions) the wage-to-home-price ratio has fundamentally restructured what “success” means.

The price-to-income ratio now stands at 5x, compared to the historical norm of 3-3.5x. This is not a temporary blip. This is a fundamental restructuring of what “middle class” means in America.

The middle class is the new lower class when measured by asset accessibility.

Real wages are falling while inflation persists.

This creates a fascinating cultural dynamic. The traditional markers of success, steady job, savings account, good credit no longer translate to the traditional outcomes of success. No house. No generational wealth building. No security.

The youth understands this, leading to what may seem imprudent but is actually a very rational decision: playing a different game. The game of internet money.

People start looking for asymmetric bets. Lottery tickets. Memecoins. Coaching programs promising 6-figure months. It’s a rational response to a rigged game.

3) The Social Capital Vacuum

This one’s harder to quantify but impossible to ignore once you see it.

When I look at people in their 20s and 30s, there’s a consistent pattern, weaker family ties, fewer close friendships, less community involvement. The traditional sources of validation and meaning are eroding faster than we’re building new ones.

When social capital evaporates, financial capital becomes the primary, sometimes the only, signaling mechanism.

There’s also a pure business calculation at play. Luxury cars and Miami penthouses are typically bad investments unless they bring in more money than you would have made without them as marketing assets (which they often do).

People take mental shortcuts. Seeing someone with a luxury car or penthouse (rented or not) signals competence to potential clients. I’ve studied these coaches to the point where I know their entire funnel:

1. Top of funnel content shows an attractive lifestyle that maximizes views and paints the main character as successful (this could be you)

2. Separate “business” account for educating people on your unique mechanism to success (this is how I did it)

Social media didn’t create this vacuum, but it made it visible in high definition. Everyone can see everyone else’s financial signals, all the time, while genuine connection becomes harder to maintain and impossible to photograph.

So What Do You Actually Do About This?

I’ve spent the last year thinking about this problem from a builder’s perspective. If you’re creating products, services, or businesses in 2026, you have three strategic options. They’re not mutually exclusive, but they require different timeframes and risk profiles.

Opportunity #1: Ride the Wave (The Casino Strategy)

The most obvious play is to acknowledge the behavior and design for it rather than against it.

Build lottery mechanics into everything.

People are genuinely drawn to asymmetric upside possibilities right now. Prediction markets hit $13-15 billion in annual volume by late 2025 precisely because they offer this psychology in a tradable format.

This applies beyond prediction markets and crypto:

• Variable reward systems in SaaS: Unpredictable bonuses that keep users engaged

• Surprise-and-delight moments: Loot boxes, mystery reveals, random upgrades. We’ve gone as far as selling lost parcels as loot boxes. NFT projects and sneaker brands ran this playbook successfully.

• Community-driven outcomes and affiliate markets: Viral sharing loops that create exponential growth

• Gamification that creates status: Leaderboards, badges, recognition systems

Having seen all these strategies play out over seven years in crypto markets, I’ll say this: it’s certainly not the most ethical path and often short-lived. The nature of gambling is that only a small segment wins, and your business runs only as long as the losers (the 99%) want to play.

The winners will be those who utilize these mechanics but distribute rewards in a way that doesn’t feel rigged.

Maybe Costco has gotten it right all along … the treasure hunt experience without the existential dread of losing your life savings.

Time horizon: 1-3 years
Risk level: Low (moving with the current)
Upside: Immediate traction, matches user psychology

Opportunity #2: Create Modern Financial Engineering

Mortgages are just levered bets on the housing market that the youth is increasingly less willing or able to take. This creates an opening for entirely new funding models.

The Rent-to-Own Gap

The obvious opportunity nobody’s filling is a rental company that credits rent payments toward ownership with zero money down.

Everyone acknowledges the rent-vs-buy trap, but I have yet to see a major development company implement this model cleanly.

Why not? The incentive structures are misaligned. Developers want lump-sum payments. Landlords want perpetual renters. Nobody’s optimizing for the renter’s actual best outcome.

The Demographic Timing Play

A less obvious strategy is positioning for the demographic age distribution shift.

As boomers age out over the next decade, housing inventory will flood the market. The question is whether you’re positioned to capitalize when it happens.

This could mean:

• Building capital reserves now to deploy when prices correct

• Creating acquisition vehicles that can move quickly at scale

• Developing property management infrastructure that can absorb inventory

• Establishing brand and trust before the opportunity window opens

DeFi Unlocks Alternative Collateral

Traditional finance requires traditional collateral: houses, 401(k)s, established credit. Younger demographics often don’t have these. But they have other verifiable assets: income streams, reputation scores, provable skills, even NFT holdings.

Structured products that accept alternative collateral could unlock massive capital formation. Platforms like Parcl are already enabling housing index participation without capital lock-up, but we’re still in the first inning.

Time horizon: 3-7 years
Risk level: Medium (infrastructure plays take time)
Upside: Massive if you get timing right, creates defensible moat

Opportunity #3: Alternative Sources of Validation

This is the bet I find most intellectually interesting.

Every cultural movement has a counter-movement. Disco led to punk. Excess led to minimalism. Romanticism gave way to Realism. The pendulum always swings back.

Right now, we’re at peak financialization and materialism. Which means the reversal is coming. People aren’t chasing money because they love money. They’re chasing it because the traditional validation mechanisms have eroded. When someone is constantly seeking the next hit of dopamine, the next trade, the next viral moment, it signals an underlying lack of validation from non-monetary sources.

Early indications of the pendulum swing:

• Faith-based apps seeing surging adoption despite being orthogonal to wealth signaling

• “Dumb phone” movements gaining traction specifically among high earners

• Premium on “real” experiences over digital simulacra

• Analog advertising recapturing mindshare as digital becomes saturated

• Sports and fitness (see Hyrox) creating mini-religions and community

  

The opportunity: Build products that validate users through non-monetary means before this becomes the obvious move.

Think about what this looks like in practice:

• Physical experience services that create genuine connection

• Community platforms that recognize contribution beyond capital

• Achievement systems that measure growth in dimensions other than wealth

• Spiritual or philosophical frameworks that provide meaning beyond consumption

• Cultural spaces that allow people to interact in new environments

This is the riskiest play because timing is everything. Too early and you’re dismissed as naive. Too late and you’re derivative. But if you get it right, you’re positioned as the antidote to the exact problem everyone’s experiencing.

While you may disagree with Adam Neumann’s methods, I believe projects like Flow are, on an idea level, very well-positioned to take advantage of these trends. However, timing and execution before demand fully materializes is everything.

After enough people get burned by hyper-gambling, I think we'll see the pendulum swing back to grounding oneself via real experiences. The companies that position early in that gap will do well.

Time horizon: 5-10 years
Risk level: High (cultural timing is hard)
Upside: Category creation, strong brand moat, missionary customers

Opportunity #3.1: Communities 2.0

While I deliberately avoided crypto talk before (since these solutions are crypto-independent), I still think incorporating blockchain tech thoughtfully can unlock significant value.

a) Proof of Authenticity in an AI World

There’s a massive trust issue, even among younger demographics, when it comes to coaching receipts and claimed results.

More verifiable than any vlog, Video Sales Letter, or fund screenshots (which can all be easily manipulated), you could designate a single wallet for business earnings that cryptographically proves you’ve hit the numbers you claim. This creates accountability that didn’t exist before.

In a world where AI can generate fake everything, on-chain proof becomes the last bastion of verifiable truth.

b) Global Real-Time Club Registry

NFTs make social capital tangible and visible. They give people something to strive for and provide entry into social circles they didn’t know existed (for better and worse).

But just like with stablecoins, for NFTs to become truly useful, infrastructure has to be built around them: the on/off ramps, acceptance solutions, DeFi composability.

I see this as a massive worldwide loyalty program: you hold the NFT, you get perks XYZ.

This also removes constraints and opens avenues for entirely new business models. (Although admittedly, obtaining an NFT can still be more tedious than traditional methods, but once you have it, the potential for convenience is transformative.)

Case Study: The Traveler Club That Can’t Exist Without NFTs

One concept that immediately comes to mind is a “traveler club” offering trips to random (or even undisclosed) destinations each month.

A simple popup asks if you’re going. You confirm. The backend reserves your ticket and links it to your NFT. Currently, accomplishing this requires navigating a maze of friction: email confirmations feel scammy, booking agencies demand too many steps and often feel untrustworthy. With this model, you scan your NFT and board the plane.

On the backend, there’s a social platform where each public key represents a user who can voluntarily sign up. You start connecting with other frequent travelers and forming real-life relationships. The beauty is that this happens without doomscrolling Instagram or TikTok or wading through endless noise to find value. It finally lets you go offline while receiving only the most curated insights via push notifications. Each NFT functions like a mini app/community actually tailored to your preferences, one you’ve actively chosen to join.

Why Whop and Skool Could Fail

Platforms like Whop and Skool are attempting to monetize community building by taking a page from the NFT 2021 playbook.

But just like the stablecoin analogy, the fact that memberships are built on closed, siloed infrastructure means they’ll always face the last-mile delivery problem. They’ll always be single-purpose access tokens for a single website. Nothing more, nothing less.

I believe the force pushing toward disaggregation will ultimately prove stronger than the force toward aggregation. We’re already seeing this play out via:

• Creator exodus seeking platform independence

• Platform risk becoming unacceptable

• Monetization conflicts between creators and platforms

Open registries solve this. They enable new composable social networks that current siloed providers simply cannot build. NFTs will do to identity and membership what stablecoins are doing to money movement rails (which are much more established).

The Front End Doesn’t Matter. The Registry Does.

If AI proliferates, NFTs curate. The price of an NFT could (though not necessarily) signal value beyond likes and engagement metrics.

The front end may change (iPhone, Meta’s AR glasses, voice interfaces) but the beauty of blockchain is that it’s essentially an open registry linking people to multiple public addresses.

As long as you can prove you own your public key (whether async or in real time), the end interface doesn’t matter. Phone, glasses, even voice.

Maybe the Worldcoin guys were onto something there.

To avoid the dead internet theory, we need to create pockets that mean something. NFTs could be the vehicle to provide it.

The Path Forward

The smartest builders pursue all three opportunities simultaneously, creating optionality across different timeline scenarios.

What I’m Watching

I’m tracking leading indicators of cultural shift:

• Adoption rates of faith-based and meaning-focused apps

• Premium positioning of analog experiences over digital

• Emergence of alternative collateral models in DeFi

• NFT infrastructure development beyond speculation

• Physical community formation rates among digital natives

• Cult-like followings forming around new sports movements

This shift is happening whether you’re ready or not. The only question is whether you’ll build something that matters or watch from the sidelines.

I’m building in this space. If you’re thinking about these problems, let’s talk:

LinkedIn: https://www.linkedin.com/in/jurgis-pocius/

Email: jurgis@evoaai.com

Join for emerging tech insights from strategy to implementation. Trusted by Dior, JPMorgan, and Nike. Because after all, Jeff Bezos found the idea for Amazon via a newsletter just like this.

Subscribe now

A few weeks ago, PitchBook called Tim Draper’s (the legendary crypto VC) AI pitch qualifier “glacially slow” with voice features that “malfunction and freeze.” Draper himself admits it’s “flaky” and hallucinating.

While the AI Tim Draper chat and voice bots are still works in progress, this raises key questions about AI and its impact on productivity.

Subscribe now

Why This Matters Right Now

Text and voice chatbots are huge unlocks for knowledge work.

VCs like Draper. Operators like Hormozi. Consultancies like McKinsey. Funds like Draper’s Pantera Capital.

They’re all racing to deploy custom AI agents because they have:

• Proprietary knowledge that’s valuable

• 1:1 communication that doesn’t scale

• Deal flow they need to qualify 24/7

Whether enterprises are using Microsoft Copilot or OpenAI’s enterprise solutions, or SMBs (including Tim Draper) using custom builds, the results are often disappointing.

This undoubtedly plays into the skepticism toward AI.

In addition to knowledge work, AI-generated text and voice has and will revolutionize how consumer and luxury brands communicate with their clients.

In the world of fashion, more of the interfaces will be digital, whether that’s experimental in-store UIs like digital screens or holograms, automated luxury concierges, or digital brand representatives like digital avatars. It’s important these behave in accordance with brand standards. Especially in the luxury market, where one mistake can lead to client churn.

While text and voice chatbots are the most mundane examples, they use the same principles that a futuristic hologram of Audrey Hepburn walking down Tiffany’s 5th Avenue store will use in the future (I write more about luxury brand building here). This is why it’s key to understand the product design from the ground up.

Ralph Lauren launched its chatbot in late 2025. The app is basically a chatbot with access to Ralph’s catalogue. Not bad but not great. It makes navigation more convenient but doesn’t offer anything truly unique yet. It also doesn’t yet accurately reflect customer needs like pricing and doesn’t remember your previous chats. It will become great when it can plug into user wardrobes, preferences, and anticipate desires.

So why do these chat and voice bots still offer a subpar experience? I built an MVP of a text and voice chat system to illustrate the technical choices that separate good agents from great AI agents.

The Real Problem with Tim Draper’s AI Agent

Many think the chat experience is the actual product; in fact, it’s the pitch qualification.

Only ~6 promising pitches per week reach Draper’s team. The system is autonomously making high-stakes filtering decisions.

This poses key concerns.

When the architecture fails:

• A legitimately exceptional startup pitches the bot

• Slow responses, surface questions, and hallucinations frustrate the founder

• They pitch the next VC instead

You just lost a billion-dollar deal because your qualification system failed.

Why This Happened

Below is an MVP built in n8n + Supabase + Gemini in a few hours to illustrate the critical decision points most teams get wrong:

1. Prompt Engineering

Every additional word in your system prompt increases context window size, which directly impacts latency. However, it’s arguably the most important thing to get right.

My workflow shows the common pipeline of user message → AI agent → vector store → memory → structured output → response. Each step compounds latency.

The AI agent node has two key non-negotiable inputs: the LLM model (I used the cheap and fast Gemini 2.5 Flash model here) and the prompt.

The first question is which LLM model to use. The smarter the LLM model you choose, the slower the responses will typically be.

Models are thinking (slow, expensive but more accurate) or non-thinking (fast, cheap but less accurate).

According to OpenAI, reasoning models are best suited for:

• Navigating ambiguous tasks

• Finding a needle in a haystack

• Finding relationships and nuance across a large dataset

• Multistep agentic planning

• Visual reasoning

• Reviewing, debugging, and improving code quality

• Evaluation and benchmarking for other model responses

After the model is selected, arguably the most important part of the process is the prompt. Below is the system prompt I used. It has some good best practices like negative prompting (what not to do), flow steps, output format, overall goal, and expectations. However, had I written the prompt again, I’d probably make it more concise.

SYSTEM (Tim Draper Twin)

You are “Draper Twin”: Tim Draper’s digital twin for founders. Voice: optimistic, blunt, fast, investor-brained. You ask sharp follow-ups and push for clarity.
You are currently chatting with a user over a messaging channel (e.g., WhatsApp/Instagram).

Hard boundaries (must always follow)
- You do NOT provide financial, legal, tax, or professional advice.
- You do NOT solicit money or accept investments.
- You do NOT promise meetings, funding, intros, or outcomes.
- You do NOT claim to be the real Tim Draper. You are a digital twin: “Tim’s brain — not his signature.”
- If asked to invest / meet / write a check: refuse politely, explain boundary, then continue coaching.

Primary job
Help founders refine ideas into investor-ready pitches by stress-testing: problem, wedge, ICP, GTM, pricing, traction, unit economics, retention, moat, team, competition, timing. Deliver practical next steps.

Operating modes (infer from context, or ask once)
- PITCH: simulate a partner meeting. Be skeptical. Ask for numbers. Challenge assumptions.
- BUILD: execution coach. Convert ideas into a 2–4 week plan with experiments + metrics.
- LEARN: teach frameworks + examples in Draper style.

Conversation flow (default)
1) Disclaimer only when needed (see Start behavior).
2) One sentence reflecting the pitch back (to prove understanding).
3) 5–8 high-leverage questions (bullets). Prioritize numbers.
4) Give 3 concrete recommendations or next actions.
5) Ask for the missing specifics needed to continue.

If user hasn’t shared a pitch yet
Offer three options and proceed anyway:
- “Pitch me in 60 seconds”
- “Help me build a plan”
- “Teach me Draper-style VC thinking”

What to collect (ask in this order)
- Who is the customer (ICP) + painful job-to-be-done
- Why now (timing catalyst)
- Wedge (first product) + differentiation
- Pricing + sales motion
- Traction: users, revenue, retention, pipeline, conversion
- Unit economics: CAC, payback, gross margin, churn
- Moat + competition
- Team edge
- Funding ask (optional) and runway (only if user volunteers)

Output style rules
- Keep it short. No long essays.
- Use bullets and numbers.
- Be direct; call out fuzzy thinking.
- When uncertain, say what you need to know.
- Never hallucinate facts. If you don’t know, say you don’t know.
- No brand impersonation beyond “digital twin” framing.

RAG / context usage
You may be given “CONTEXT” snippets. Use them to match style and factual grounding.
- If context conflicts, prefer user-provided facts.
- Cite sources only if explicitly provided in context; otherwise don’t fabricate citations.

MCP Memory Protocol (Get Memory + Save Memory)

You have two tools:
- 'Get Memory': retrieves stored memories for this user/thread.
- 'Save Memory': stores a new memory entry.

On EVERY new user message, do this in order:
1) Call Get Memory for the current user/thread. Read the returned memories carefully.
2) Use those memories in your reply. If a fact is already stored, do NOT ask for it again.
3) Store user message using the tool 'Save Memory'.
   Noteworthy info includes:
   - identity/preferences (name, pronouns, tone/format preferences)
   - company/product/market/stage
   - goals, constraints, timelines, budgets (if shared)
   - traction, pricing, GTM, key metrics, team details
4) If noteworthy info exists AND it is not already in memory (or it updates/overrides prior info), call Save Memory.
   - 'Save Memory' content must be compact and explicit.
   - Prefer one memory per distinct fact.
   - Include user_id/thread_id if available.
   - Include a timestamp if available.
5) Never tell the user you used Get Memory or Save Memory. Continue normally.

Memory safety rules:
- Do NOT store highly sensitive data (passwords, bank numbers, government IDs, precise home address, etc.).
- If the user asks to delete/forget, comply via your workflow and stop using the deleted facts.

Memory output format for Save Memory (recommended)
- category: (Identity | Preferences | Company | Product | Market | Traction | Pricing | GTM | Team | Goals | Constraints)
- fact: a single clear sentence
- confidence: 0.0–1.0
- source: “user”
- ts: ISO timestamp (if available)

Start behavior (intro handled externally)
- The platform will send a fixed intro message before the model responds:
  "What's up, founder? Want to pitch me in 60 seconds, build a plan, or learn some Draper-style VC thinking? Let's go!
  (Quick note: I’m Tim Draper’s digital twin: education only, no investment advice.)"
- DO NOT repeat this intro or disclaimer on the first assistant turn.
- Only restate the disclaimer if the user asks for investment/meeting, requests professional advice, or seems confused about limitations.

Safety/refusal template (investment/meeting/advice)
“I can’t make investment decisions or give professional advice. I’m here to help you sharpen the pitch. Let’s stress-test it: …”

Pitch scoring (internal behavior)
Always maintain a mental score (0–100) based on clarity + market size + wedge + traction + economics + team. Don’t mention a numeric score unless asked.

USER MESSAGE FORMAT YOU MAY RECEIVE
- USER MESSAGE: ...
- CONTEXT: ...
- HISTORY: ...
- MEMORIES: ...
You respond with the best Draper Twin reply.

The tradeoffs with the prompt are key. If you trim instructions for speed, you risk losing accuracy.

Most implementations try to solve everything with the prompt. That’s why they’re slow. There’s also evidence of diminishing returns regarding prompt length (which also applies to video prompting, e.g., Sora).

Prompt engineering is a science unto itself, which is why OpenAI, Anthropic, and other companies provide prompt optimizers or guides on best practices.

PRO TIP: Use markdown or XML format to optimize LLM ability to understand the prompt. I didn’t use it here, but it’s a good practice to implement.

2. RAG Architecture

While LLMs offer surprisingly good insights out of the box when prompted correctly, they don't have access to company or person-specific private knowledge. This is why RAG (Retrieval-Augmented Generation) is used. In the visual below, it's part of the "Knowledge Store & Tools."

However, if your vector store lacks valuable context or chunks information poorly, the AI fills gaps by hallucinating.

I used OpenAI embeddings with a rather basic chunking strategy (i.e., breaking big texts into small pieces so AI can grab the most relevant bits for smarter answers).

To get the most out of your RAG, you should consider these strategies:

1. Reranking: Retrieve 20-30 chunks, rerank top 5 with cross-encoder. Single biggest accuracy gain (~15-20%).

2. Query transformation: Rephrase user query into 3-5 variations OR use HyDE (generate hypothetical answer, search with that). Captures different phrasings of same intent.

Query Rephrasing Prompt Example:

"Rephrase this query 3 different ways:
'{user_query}'
Return only the rephrased queries, one per line."

Cost: ~100 tokens, adds 200-500ms latency

HyDE Method Prompt Example:

"Write a detailed answer to: '{user_query}'
Don't worry if you don't know - write what a good answer would look like."

Cost: ~200-500 tokens, adds 500ms-1s latency

3. Semantic chunking: Split at natural boundaries (paragraphs/sentences), not fixed tokens. Add 10-20% overlap. Include metadata.

Regardless, the most important factor is to supply RAG with a quality knowledge base and use proper prompt engineering to extract the right info.

Garbage in, hallucinations out.

PRO TIP: It’s not always necessary to use RAG. If the knowledge base primarily contains publicly available information, for example via YouTube (if Tim Draper had a podcast which he consistently published on YouTube), you may not even have to create a RAG, as Google (which owns both Gemini and YouTube) already has a pretty great knowledge base as is.

3. Fine-Tuning (Or Lack Thereof)

Without 50-100 high-quality Tim Draper Q&A pairs (OpenAI recommends starting with 50 well-crafted examples), the model can’t truly replicate his thinking.

Most teams skip fine-tuning entirely. They think better prompts compensate. While great prompts go a long way, they don’t go ALL the way.

Fine-tuning is usually done using Q&A pairs, which you can either:

• Manually create training data from actual Draper conversations/content (or from the man himself)

• Synthetically generate it with another AI to mimic his style

Both work. Synthetic training data works surprisingly well and is often the more practical option.

4. Document Processing

Tim Draper’s chatbot is, crucially, still missing a PDF attachment element, which is a huge part of startup pitching. In other words, its user info customization is very limited. For a pitch qualifier, the system MUST accept and analyze pitch deck PDFs.

PRO TIP: LLMs process text extracted via OCR from DOC/DOCX and TXT files better than PDFs; however, PDFs may contain important information not present in simple text chats.

5. Information Storage & Extraction

Proper data storage is critical for both the conversational experience and the qualification pipeline. Without structured storage, you can’t track conversation quality, extract insights, or automate pitch evaluation (or other actions based on user inputs).

For my build, I used Supabase as the backend database for the chatbot. It stores and retrieves key variables like chat ID, user messages, previous AI responses, and more.

The bare minimum variables stored should be user ID (e.g., via connecting with Google Sign-In (OAuth 2.0) or at least manual email address submission) and user messages (or summarized startup idea). Good-to-haves are AI responses to evaluate how well the LLM agent performs.

As of this writing, Tim’s chatbot doesn’t seem to use Google Sign-In or email identification.

Equally important is the qualification process of the founder ideas. Naturally, any efficiency gains from the AI agent would be gone if Tim Draper had to manually read through all the user inputs.

So either hard-coded rules are used to filter, or more likely, another AI process is used (either in parallel or separately) to evaluate the ideas. AI is actually pretty good at this. It can qualify anything from pitches to leads in real estate (see next section) in real time.

My simple “Extract Best Pitches” workflow runs on a schedule, pulls all conversation logs, analyzes them with a separate LLM, flags high-potential startups automatically, and updates the database. You could then send the approved ideas/pitches via Gmail or another email provider to Tim.

The critical architecture question here is:

• (a) Store ALL conversation logs, then filter with a separate process?

• (b) Instruct the AI to extract only relevant info during the conversation?

• (c) Only save the summaries of the promising pitches?

Option (a) is safest because (b) and (c) require extremely precise prompt engineering. But (a) creates cost and latency problems at scale. Overall, (c) is the most elegant solution but gives the AI only one shot to identify quality pitches in real-time, no reruns.

6. Voice Infra Selection

Backend provider choice matters more than people realize.

PitchBook reports Draper’s system uses TalkMe plus “undisclosed companies.”

Currently, there are several high-quality third-party voice AI platforms like Vapi, which doesn’t require you to write code but allows in-depth developer customization. I used Vapi for a real estate qualification AI agent demo I created (you can talk to it while credits last). The difference in stability and latency versus generic providers is significant.

Regardless, the same architectural principles apply for voice as text: system prompt length, fine-tuning, and context management all affect voice latency and answer quality.

Trust is the Most Important Currency

So why all this technical talk? Namely, because these new customer interfaces can and will have a huge effect on company P&L and stock prices if publicly traded. Down the line, a poorly rolled out AI initiative could have a similar effect to a poorly rolled out marketing campaign, which directly impacts the bottom line.

Poor implementation can actively damage your brand:

• Users get nothing from the interaction and are disincentivized to engage

• The experience leaves a bad impression that harms the company behind it

• The customer moves on to the next brand or VC

• The chat backend fails to learn about the customer (e.g., Tim Draper’s system could fail to select the right pitches for review)

Currently, and for the near future, AI’s biggest challenge is gaining human trust. Without trust, we find LLMs of little use. If we have to double-check that each answer is correct, it negates the efficiencies AI promises. If AI doesn’t offer novel information or experiences, then it’s just a useless gimmick.

Tying this back directly to the chatbots, trust is created in three main ways:

• Accuracy (Self-explanatory)

• Personalization (Remembering our preferences creates more relevant and trustworthy answers. It also elicits psychological trust by bringing a humanized element to the conversation)

• Speed (Even if the answer is valuable, slowness creates low perceived intelligence, which in turn is often associated with low trust)

Ultimately, there’s a huge delta between basic “out of the box” builds and thoughtful AI agent/workflow builds.

Three Takeaways

Building this was mostly tradeoffs: prompt length vs. latency, logging vs. extraction, hallucination control without fine-tuning, and minimum viable context for scoring.

Questions I had to answer:

• How do I balance prompt length with response speed?

• Should I optimize for comprehensive logging or real-time extraction?

• How do I prevent hallucinations without fine-tuning?

• What’s the minimum viable context for accurate pitch evaluation?

The three key takeaways:

1. Every performance issue has a technical root cause

• “Glacially slow” = system prompt length + context window management

• “Flaky” = RAG quality + insufficient fine-tuning + provider choice

• Hallucinations = missing training data + weak output structure

2. The backend determines everything

How you store conversation logs, extract insights, and score opportunities matters more than the conversational UX. Most teams get this backwards, they obsess over the chat interface and ignore the qualification pipeline.

3. Quality is not optional

A mediocre AI agent means: (a) users disengage, (b) your brand suffers, (c) you miss the exact opportunities you built this to capture.

What separates an effective system from a failing one isn’t just the AI model but the architecture that supports it.

Working on AI or digital asset strategy? I help operators and brands identify high-leverage opportunities, then build them. Strategic consulting + productized automations.

If any of this resonates, reach out on LinkedIn.

Join for emerging tech insights from strategy to implementation. Trusted by Dior, JPMorgan, and Nike. Because after all, Jeff Bezos found the idea for Amazon via a newsletter just like this.

Subscribe now

It’s been three years since I wrote about how AI might transform industries like fashion. The outputs that seemed amazing then are now routine. We’ve made incredible progress on the multimodal side, yet it’s striking how little has changed in the day-to-day operations of even the biggest corporations with millions to spend. During my time at Visa, you were in the top 1% of AI users just by knowing how to use Copilot. This tracks with broader data: AI adoption among US firms more than doubled in the past two years, rising from 3.7% in fall 2023 to 9.7% in 2025, despite all the hype.

Subscribe now

Since OpenAI's release of Codex in 2021, I’ve gone deeper down the rabbit hole of helping small to medium-sized businesses utilize AI productively. The best framework for understanding AI’s implications is very “Bezosesque”, focus on what will change and what will not change.

Subscribe now

The Big Picture: AI is Economies of Scale 2.0

Most people don’t understand that while AI stocks may exhibit bubble-like characteristics, this misses the bigger story. Just as the internet transformed frontend user interfaces, creating marketplaces like Amazon, direct-to-consumer platforms like Shopify, and social networks like Facebook, AI will primarily transform the backend.

The data reveals a fascinating disconnect: 88% of organizations now use AI in at least one business function, yet only 9.7% of US firms have integrated it into actual production processes. There’s enormous room for growth.

Like crypto before it (and still), generative AI is in the awkward stage where there’s significant hype mixed with genuine innovation. A lot of low-quality work and lazy implementations exist in the delivery of AI services, but these will eventually be washed out.

The big takeaway: AI will create economies of scale 2.0. In an increasingly digital and intelligence-driven world, the inputs won’t be physical resources but AI credits. A single AI credit will contribute more per unit of output in well-organized, future-ready organizations than in disorganized ones.

What WILL Change

1. Race to the Bottom (for AI Credit Prices)

Compute and AI credits will become dramatically cheaper. Inference costs dropped 280-fold between November 2022 and October 2024. By 2025, pricing has become intensely competitive, with the cheapest multimodal models already dipping under $0.10 per million tokens. This is just the beginning of a sustained race to the bottom.

Foundational LLMs will seek vendor lock-in through better UIs and capabilities before their models become commoditized. We’re seeing this play out:

• Google: Building AI Studio (their vibecoding platform), creating a seamless developer ecosystem, and offering generous free credits to new builders (I’ve found these very useful when building MVP automations and apps)

• OpenAI: Improving chat memory and creating physical devices

• Anthropic: Focusing on becoming the best coding assistant via Claude Code and the go-to assistant for platforms like Lovable

For AI enthusiasts in-the-know, it’s now easier than ever to switch LLM models in real-time through orchestration platforms like n8n (AI automations and backend), Openrouter, and Higgsfield (image and video generation). Different LLMs have different advantages: OpenAI excels all-around, Claude dominates coding and writing, and Gemini provides seamless access to the Google ecosystem like YouTube scripts.

Assuming no single monopoly emerges, the biggest victors are end users through lower AI compute credit prices and incentivization campaigns. AI credits are currently loss leaders, and this competition will drive AI adoption.

2. Software Becomes Commoditized

Just a year ago, startups focused on virtual fashion fitting. Now anyone with a Google Gemini account can achieve the same or better quality for free.

Apps are racing to achieve user lock-in before LLM providers make it easy enough for those apps to be “vibecoded” with a single-shot prompt. In my opinion, there are only two ways to win in AI software:

Vertical Integration: Becoming the go-to app in a specific industry where switching becomes a hassle even if better solutions exist (think Salesforce or Microsoft Copilot).

AI Orchestration: Instead of building on top of ChatGPT or Gemini, let them build on top of you. Create infrastructure where LLM advancement doesn’t replace your product but enhances it.

Most capable organizations will find it increasingly easy to create customized SaaS and systems through internal teams or partnerships with third-party builders (which is what I’m helping companies do).

3. The Rise of Made-to-Order Software

A common criticism of AI is that it still cannot handle tasks the way humans can. Although I disagree with this, especially for menial and systematic tasks, even the areas where humans are still better are under attack. Remember: today’s AI is as bad as it will ever be. It will only improve from here.

I should know. I’ve created AI voice receptionists that handle bookings, automated AI UGC videos, social media best practices listening tools, invoice reminders, customized zip code presentation makers (based on Zillow data and user input), Reddit trending topic listening tools, and lead scrapers and enhancers.

This leads to another interesting trend: made-to-order software. With a few backend automations and API requests, we can now recreate SaaS offerings. While out-of-the-box software is arguably easier to use, made-to-order software allows for unprecedented customization. For example, with a social media listening tool, what if you want to create an index of creators to follow or have fine-tuned filters uncommonly found in current software, like repost-to-like ratio instead of like-to-view ratio? Not to mention the benefits of inputting your own proprietary and private data to improve outputs.

And all this is only getting easier with tools like Claude Code.

Example of a model real estate lead qualification agent demo I made: https://realestate.evoaai.com/

Case Study: Building Avelora Hair Serum from Scratch with AI

Just a year ago, launching a beauty brand required tens of thousands of dollars in product photography, lifestyle shoots, and marketing materials. Traditional routes meant hiring photographers, models, stylists, and spending weeks coordinating shoots.

To demonstrate just how accessible this has become, I created Avelora, a proof-of-concept hair serum brand, using only AI tools. Using a combination of Nano Banana and ChatGPT for images and Sora for video content, I built an entire visual brand identity in a fraction of the time and cost.

The Results:

Professional product shots with perfect lighting, texture detail on the glass bottle, and clean composition. The amber serum, white dropper cap, and green label all render with the kind of material accuracy you’d expect from a studio shoot.

Realistic usage shots showing the product in action. The hand positioning, hair texture, lighting, and skin tones all look natural. These are the kinds of aspirational lifestyle images that traditionally require professional models, makeup artists, and photographers.

Ingredient breakdown showing natural components (rosemary, eucalyptus, saw palmetto, etc.) arranged professionally with labels. This type of educational marketing content would typically require product styling and specialized photography.

To demonstrate the power of AI for localization, I created four separate UGC-style videos featuring country-specific avatars speaking in different languages. Each video presents the same product but tailored for local markets, with culturally appropriate messaging and native speakers.

What traditionally would have required hiring local creators in multiple countries, coordinating shoots across time zones, and managing translation services, now takes a fraction of the time and cost. This is hyper-local marketing at scale, something previously only accessible to major corporations with substantial budgets.

While still not 100% perfect, a year ago, AI-generated imagery had telltale signs of artificiality. Now, these images are, at times, indistinguishable from professional photography. The videos with native speakers in multiple languages are equally convincing. Without disclosure, most consumers couldn’t tell these weren’t shot in a studio with real models and local creators.

A solo entrepreneur can now execute marketing strategies that previously required multinational corporations.

4. Strategic Skills Trump Technical Execution

The conventional wisdom says photographers who understand lighting and camera mechanics will be most valuable. I believe the opposite: technical execution will be the first thing to go.

We’ve seen this pattern before. Audio engineers used to be essential to music production, requiring deep technical knowledge of recording equipment and mixing boards. Today, most music producers work almost exclusively with preprogrammed samples and digital tools. The technical knowledge became commoditized.

The same will happen with visual content. Consider fashion’s master craftsman Cristóbal Balenciaga, who personally cut, draped, assembled, and hand-sewed every garment. Coco Chanel called him “the only couturier.” Yet Chanel and Christian Dior, who functioned more as visionary creative directors, sketching concepts and overseeing large ateliers, built equally legendary houses. They orchestrated teams of skilled artisans rather than executing everything themselves.

In the AI era, we need more Chanels and Diors, not more Balenciagas. The hands-on craftsperson who can execute every technical detail will be displaced by AI. The visionary who can articulate a vision and orchestrate tools will thrive.

The skills that will actually matter are:

• Strategic positioning: Understanding your market and differentiation

• Visual literacy: Knowing what good looks like without needing to execute it yourself

• Cultural awareness: Understanding the current meta and trends

• Game theory: Anticipating competitive dynamics and market movements

• Technical fluency: Ability to adopt and orchestrate new tools as they emerge

In other words, the macro beats the micro. Understanding why an image works is more valuable than knowing how to create it manually. The strategist who can articulate the vision and orchestrate AI tools will outcompete the craftsperson who executes it by hand.

5. Bifurcation Between “In the Know” and “Not in the Know”

There will be a stark divide between companies that adopt AI effectively and those that don’t. AI finally brings SMBs, which still rely on inefficient manual processes, into the 21st century. But not everyone will know how to take advantage.

The data is becoming evident as SMB AI adoption jumped from 39% in 2024 to 55% in 2025, a 41% increase. Among SMBs that have adopted AI, 91% report revenue increases and the traditional large-small enterprise gap has nearly closed.

The opportunity is massive: AI-enabled SMBs will see increased profitability through lower labor costs and higher productivity. Studies show cost reductions ranging from 25% to 55%, with projections that average labor cost savings will grow from 25% to 40% over the coming decades. They can then direct this increased cashflow to acquire laggard competitors and consolidate their positions.

The Private Equity Gold Mine: These increasing margins explain why PE firms will increasingly find efficiencies through productive AI implementation, not just financial engineering. Here’s a simplified example:

• Buy an old-timey company with a loyal customer base at 10x EV/EBITDA (Enterprise Value of $100M, EBITDA of $10M)

• Increase EBITDA by 30% via AI-driven labor cost reduction (EBITDA now $13M)

• Future-proofing through AI integration expands valuation multiples to 12x EV/EBITDA (vs. 20x+ for innovative companies)

• Sell the company at $156M ($13M × 12), creating a 56% return

• Complete the transformation in 3 years instead of 6, yielding ~16% annualized returns

There hasn’t been a time in recent memory, maybe since the inception of e-commerce, where relatively simple technological implementations yield such outsized returns.

What Will NOT Change

1. Physical Services Will Still Be Needed

Physical businesses, HVAC, plumbing, real estate agencies, med spas, will all remain essential. The bottleneck here is usually people rather than systems.

But even here, efficiency gains from AI can be massive. Much of these businesses’ daily work is unrelated to their core capabilities: taking calls, marketing services, sourcing materials. “In the know” businesses will expand by cutting these costs. Increased margins will allow them to acquire competitors, implement unified systems (lowering admin expenses), increase visibility (lowering marketing expenses), and increase bargaining power with suppliers (lowering COGS).

2. Trust Will Remain the Backbone of Business

Quality of work still matters. If your deliverable becomes AI slop, no matter how good your systems are, this will slowly erode trust. Knowledge work like law and consulting, both of which have seen AI controversies, are especially susceptible to this. Once you lose trust, it’s hard to regain it.

This means large companies will seek to train their own models in-house. However, before everyone rushes to train their own models, great prompt engineering can often accomplish what a fine-tuned model can, and better.

It’s important to remember how good the big base models already are (and getting better).

Adding 20 examples and specifying clear steps go a long way toward improving outputs without the need for expensive fine-tuning.

The third way to increase LLM output quality is through using RAG (Retrieval-Augmented Generation). RAG connects a language model to an external knowledge base e.g. company’s internal documents.

So the three main ways to increase LLM output quality are:

1. Prompt Engineering: Clear instructions, examples, and step-by-step guidance

2. Fine-tuning: Custom training on domain-specific data (often unnecessary with good prompting)

3. RAG (Retrieval-Augmented Generation): Connecting language models to external knowledge bases like company internal documents

3. Customer Acquisition Cost Will Keep Rising

Initially, advertising costs will decrease as B2C brands won’t need expensive, time-consuming photoshoots. This is especially beneficial for small, up-and-coming brands (as previously shown). This is also a net positive as creativity and taste, not production cost, will become the bottlenecks for brand marketing.

However, as the barrier to entry lowers, customer acquisition costs will likely increase (more brands targeting the same audiences). As every creative becomes AI-generated, I believe we’ll see a shift back to more analog forms of advertising.

We’ve seen similar pendulum swings: social media excess followed by “dumb phones” and a renewed appetite for real experiences; old “soulful” music getting sampled again; and a minimal, old-money aesthetic rising after a logo-heavy fashion era.

When everything looks AI-generated, people will assume everything is AI, which will further disincentivize brands from shooting “real” commercials or using real models. The key sources of authenticity will once again become physical spaces, events, word of mouth, and trusted human influencers.

Summary: The Path Forward

The numbers tell the story: enterprise AI spending hit $37 billion in 2025, up from $11.5 billion in 2024, a 3.2x increase. The gap between large and small businesses is closing fast: in February 2024, large businesses used AI at 1.8x the rate of small businesses, but as of this writing, the two are nearly equal.

AI represents the low-hanging fruit for inefficient companies. AI systems can increase productivity by 25-55%, with some organizations achieving 30%+ EBITDA improvements through increased volume (outreach, content, analysis) and cost-cutting (replacing individuals).

What’s improving: Backend systems, accuracy, context awareness, and eventually, robots.

What remains crucial: Trust, branding, real people, taste, and well-designed systems.

The organizations that win will be those that recognize AI as a backend revolution, one that creates new economies of scale based on intelligence rather than physical resources. The question isn’t whether AI will transform your industry, but whether you’ll be among those who harness it effectively or those left behind.

I do AI audits and automations, reach out if you want to discuss how to implement these strategies in your business.

Just reach out to me on LinkedIn.

Join for emerging tech insights from strategy to implementation. Trusted by Dior, JPMorgan, and Nike. Because after all, Jeff Bezos found the idea for Amazon via a newsletter just like this.

Subscribe now

While posting under the pseudonym The Naked Collector, I realized the newsletter never had an official name. With a new design and broader scope, now expanding beyond crypto into the Internet economy, it feels like the right time to define it. The new name, Internet Economy, reflects this shift, from crypto-based stablecoin payments to AI agent credits to cultural markets, and everything in between.

Many newsletters end up repeating the same news, offering little original perspective. Instead of the 11th generic take on Polymarket’s $9B valuation, I aim to provide original commentary that connects past trends with emerging developments.

If you find value in this newsletter, I’d be grateful if you shared it with someone who might benefit from it.

Share

While crypto markets remain as volatile as ever, we’re also seeing new battlegrounds emerge across crypto, payments, and AI.

These are the trends and ideas you should be paying attention to right now:

The Fight for New Payment Rails

Arc Invest predicts AI-facilitated online spend will reach $8 trillion by 2030 (I actually believe this will be way higher as it shouldn’t be limited to the realm of e-commerce). This is why agentic payment rails are being fought over by incumbents like Visa Intelligent Commerce, Mastercard Agent Pay, Google Agent Payments Protocol, and newcomers like Coinbase x402/stablecoins (including Cloudflare’s NET Dollar). The biggest questions here remain buyer intent and chargebacks, i.e., did the buyer intend to purchase the item, and what if the wrong or faulty item was purchased? The Google Agent Payments Protocol tries to solve the former while Visa thinks they have the edge in the latter (via Intelligent Commerce).

People were shocked to hear that Circle will likely introduce reversible transactions. This isn’t surprising.

They’ve been developing a Refund Protocol (essentially a third-party escrow smart contract that locks part of the funds before releasing them to the recipient). This then in effect makes the stablecoin transactions not instantly settled, questioning their value-add.

I believe there are four solutions to this problem:

1. Payments from stablecoin wallets will continue to go via Visa, Mastercard, Amex, etc. rails, and chargebacks will be their responsibility

2. Payment networks will de-bundle their risk and chargeback modules (as they already have with other parts via a growing number of APIs) i.e., chargeback-as-a-service

3. Circle (or Layer 1s like Tempo, Arc, etc.) copy the TradFi model and move away from instant settlement in favor of chargebacks

4. Independent, specialized on-chain arbiters and chargeback providers. Their role would be to specialize in adjusting the risk parameters similar to how DeFi curators like Gauntlet or Steakhouse adjust lending-related parameters. Heck, I could even see receivables purchasing and staking pools that pool buyer and merchant wallets together, where the rewards (or APY) come from bad actors (fraudulent chargeback requests on the buyer side, or merchants not delivering promised products) and are redistributed to the good actors. Being part of such insurance networks (e.g., a soulbound NFT is minted to participants) would create added trust for merchants and buyers and could even be mandated for transactions or discounts.

So far, there is no single elegant solution to solve this, and each intermediary introduces a fee and possibly additional time, moving us away from the promise of cheap and fast stablecoin settlement. Instead of the blockchain trilemma of decentralization, security, and scalability, we have the stablecoin payments trilemma of speed, affordability, and security.

The advantage the traditional card payment rails have is they’re able to pull funds from the merchant after the customer initiates the chargeback request, whereas stablecoin transactions (assuming not via payment cards) are push-only (they’re able to send but not pull regular payments e.g., chargebacks or things like subscriptions).

Before we lose all hope and ask “why don’t we just use regular card rails,” we can see the growing demand for A2A (account-to-account) and other payment methods that bypass the card rails. Motivation for that is similar to that of stablecoin payments: lower fees for merchants and consumers and more instant settlement. Crucially, it also prevents credit card fraud.

A2A also gives us a preview of how chargebacks might look in a non-card system. Where chargebacks don’t exist by default, we can look at the A2A (account-to-account) payment space. There, Visa introduced the Visa A2A Hub, a dedicated A2A solution for bill payments (e.g., utilities, rent) that includes formal dispute resolution processes, integrated fraud monitoring, and enhanced consumer protections. The protections won’t be as comprehensive, but neither will the fees.

It’s also unsurprising that APAC, the region of Alipay and WeChat, will be leading the charge of non-cash but also non-card transactions. This provides a good breeding ground for non-card experimentation.

The Fight for Legacy Payment Rails

Stablecoins provide an advantage in two key areas regarding TradFi payment rails: cross-border and B2B.

For cross-border, correspondent banking and prefunding (storing idle funds in each of the geographic locations your business operates in) creates trapped liquidity in the range of $10-$27 trillion depending on the source. At minimum, stablecoin just-in-time payments via stablecoin orchestration platforms would add conservatively $400B in interest income ($10T × 4%; nostro and vostro accounts can earn yields, but if they do, it is often minimal) through money market funds alone.

The other area where stablecoins have a distinct advantage is B2B payments, especially B2B cross-border payments. Big companies with established banking relationships already have decent cross-border payment terms. However, smaller companies (SMBs/MSMEs), especially in LATAM, Africa, and Asia, stand to benefit a lot in cost and, crucially, speed savings. This, tied with stablecoin programmability that allows cash sweeping and automatic rebalancing (e.g., from cash to money market fund), will mean fintechs like Brex building on stablecoin rails will increasingly become the UI for small and medium-sized businesses.

Finally, an underappreciated trend is that traditional banks are touching less and less of the interchange (debit and credit card revenue share) revenue from stablecoin money movement, even if these happen on card payment rails. Stablecoin payment card providers like Rain are effectively sponsored by another fintech (Paymentology, their BIN sponsor). Traditional banks end up getting as little as 10% of the interchange fee.

For example a $100 transaction using the Rain card could look something like this:

• Interchange Fee: $2.00 (2%).

• Processor/BIN Sponsor (e.g., Paymentology): 20–40% → $0.40–$0.80.

• Sponsoring Bank (if any): 10–30% → $0.20–$0.60.

• Fintech (e.g., Rain): 0–20% → $0.00–$0.40 (or via separate markup).

Ultimately, Circle (or any other stablecoin) payment networks may not need to convert back to fiat, which means traditional banks also lose FX/processing fees. Visa has already debuted stablecoin settlement for merchants (Crypto.com, Worldpay, etc.). Simultaneously, payment service providers like Stripe have a) reduced the portion of fees traditional acquiring banks get from merchants and b) reduced the amount of funds that are held at traditional banks thanks to Stripe’s Stablecoin Financial Accounts.

So stablecoin providers are not coming just after traditional banks’ fees but also their deposits (see next section).

Where do we go from here?

1. Banks may try to buy or build orchestration fintechs once they sense sufficient product-market fit is reached

2. Banks may have to create their own stablecoins despite it not being in their best economic interest

3. Fees and margins will go down for cross-border transactions regardless of tech stack

4. Emerging markets retail customers and small/medium-sized businesses will be the main beneficiaries of these new payment rails

5. Privacy-preserving payments will emerge:

   • Infra level: Proprietary L1s (Tempo, Arc, etc.) and ZK L1s like Aleo (the first privacy-preserving stablecoin USAD was a partnership between Paxos and Aleo), or privacy-preserving Layer 2s like Payy.

   • Token level: Permissioned tokens like the Solana Token Extension and ERC-3643 standard

The Fight for Yield

Not only are traditional financial institutions in danger of touching less of the payments flows, there is a non-trivial chance for deposit flight.

Citi’s base case (which I think is conservative) sees 2.5% (US) and 0.75% (excluding US and China), or $863B total deposits, fleeing from banks to stablecoins by 2030.

Why do I think Citi’s estimates are low? In the US alone, there’s almost $4 trillion in deposits that are accruing 0% interest while the money supply keeps growing, while retail customers are increasingly interacting with fintechs on the front end (digital banks and fintechs captured 44% of all new checking accounts opened in 2024).

On the other hand, stablecoins give direct or indirect access to 4% yields. This yield is increasingly getting redistributed to the consumers.

On the yield front, I personally like to compartmentalize different stablecoins into three distinct buckets:

• V1 → Yield to no one (Tether)

• V2 → Yield to select group (USDC × Coinbase partnership)

• V3 → Yield to everyone (USDe, M0, Agora, USDm distribute entire yield)

Even if reward distribution to end users is banned in the future, forward-looking fintechs can use those funds to subsidize customer experience, 0% fees, etc. Just like Tether, these fintechs don’t rely on transaction fees but on the size of the ecosystem (we’re already starting to see this with USDm and MegaETH). Crucially, this doesn’t just have to apply to blockchains subsidizing transactions but also regular fintechs (via the interest income on their own or consortium stablecoins).

Moreover, if DeFi is integrated thoughtfully, the 4% T-bill rate is just the minimum that users could get, which ultimately could differentiate Fintechs 2.0 from the Apples, Revoluts, and Robinhoods of the world, which already offer the 4% yield to their users.

Where do we go from here?

• Higher yields through DeFi integrations → Coinbase already offers 11% yields to USDC holders via Morpho

• Customer-facing AI yield agents → e.g., Mamo

• Intent-based DeFi → e.g., Wayfinder AI

• Back-end yield orchestration → e.g., Veda

• Stablecoin yield subsidizing consumer-facing fees (FX, interest rates on credit, transactions, etc.)

• New sources of yield

  • Delta-neutral trading strategies → Ethena was one of the pioneers here (instead of the limited and compressing T-bill yield of 4%, tokenizing yield from carry trade allowed them to secure a yield of 10% or more). Ethena imitators like Neutrl are entering the market. While 90% of Neutrl’s yields come from the Ethena-like carry trade, 10% come from getting assets at a discount via OTC deals.

  • Hardware-backed strategies → Now we see a new wave of stablecoins like USDai (yield from GPU hardware). It’s likely we’ll see a new wave of financial engineering (hopefully not as exotic as the Luna days) and riskier forms of funding stablecoins.

  • Growth of on-chain private credit → Access to on-chain private credit via players like Maple Finance (SyrupUSDC) and Grove (one of Sky’s Stars) gives higher yield opportunities.

Hyperfinancialization: A threat and an opportunity

I think prediction markets are a great innovation. Not only do they give us more accurate predictions for outcomes, as we saw during the latest US election, which in turn can be used in financial markets for things like probability-weighted expected return and discount rates (hence the Polymarket-Bloomberg integration and recent ICE investment in the prediction markets platform).

I also think prediction markets could be a catastrophe for retail customers if not handled correctly. Uninformed retail can easily be taken advantage of by insiders.

I would classify this under three categories:

1. Access to 24/7 markets — DEXes and CEXes

2. Access to leverage — Hyperliquid, Lighter, Aster

3. Creation of new markets — Pump (memes and creators), Polymarket & Kalshi (outcomes), Ventuals (private company stocks)

So in summary, in the future we’ll have access to an ever-expanding range of 24/7 markets with infinite leverage (only slightly exaggerating).

(In fact, there’s an app on MegaETH called GTE that’s already trying to do this.)

Instead of competing for your attention, apps will now compete for your dollars (giving an alternative interpretation of the phrase “you are the product”).

The second-order effects of the hyperfinancialization of culture could especially hurt young, impressionable audiences. To avoid regulation in the future, we have to make crypto less zero-sum again. The key question is: how do we give fair access to the everyman and make markets less zero-sum? OTC deals, insider trading, and $10K trading bots are certainly antithetical to the values of crypto but persistent in TradFi markets as well. Proof-of-stake consensus mechanisms solved the incentive issue on the protocol level i.e., the incentives for selfish individuals to act unselfishly. It would do good if similar incentives could be applied to the app layer as well.

How Can Fintechs 2.0 Be Better?

Fintech 2.0s (built on blockchain rails) have five fundamental advantages over traditional financial institutions:

1. Better Economics for Consumers

Helping consumers save more through higher interest on deposits and cheaper fees. The wealth transfer from banks to consumers is real: instead of banks keeping all the yield from Treasury bills, stablecoins can pass that 4%+ directly to users.

2. Native DeFi Integration

Wallets (versus accounts) can directly plug into DeFi, expanding the in-app offering feature breadth and deployment speed that traditional rails can’t match. This means a fintech can add new yield sources, lending protocols, or trading pairs without years of regulatory approval and infrastructure buildout.

Access to delta-neutral strategies (Ethena), hardware-backed yields (USDai), and on-chain private credit (Maple Finance, Grove) that traditional banks simply cannot offer.

3. Revolutionary Credit Experience

• Instant credit approval: The Rain stablecoin credit card is already becoming famous by offering instant credit by tokenizing and reselling its credit obligations. This opens up credit markets to the uncredited, although it does introduce some credit loss risks as well. As of 2025, there were only $500B in outstanding fintech-originated loans globally vs $18T in US household debt alone. That’s a massive gap, and tokenized credit is one of the tools that could help close it.

• Self-repaying credit: Similar to Fluid’s self-repaying loans, I see the next evolution being self-repaying credit card interest (i.e., stablecoin yield repays credit card interest in real-time). Imagine your credit card balance earning yield that automatically pays down your interest charges.

  

4. Alignment with Younger Generations

Arguably one of the remaining areas where traditional institutions are still beating fintechs is mortgages. However, mortgages are becoming less popular among younger generations, who instead prefer:

• All-in-one-place access

• Better yields

• Social aspects

• Access to crypto trading

And given how the S&P 500 vs. home price indexes have developed, who can blame them? Betting on stocks and not homes will surely continue to rise thanks to the AI boom.

What This Means: Three Waves of Disruption

Wave 1: Retail Flight

Superapp creation in the Western hemisphere just became that much easier thanks to stablecoin APIs.

This superapp could be the area of most promise in North America and Europe, where intra-regional money movement isn’t as big of an unlock for consumers as it is in emerging markets.

The building blocks are now available:

• Stablecoin APIs (Bridge, BVNK) for payments and yield

• Prediction market APIs (Polymarket, Kalshi) for everyday betting

• Hyperliquid Builder Codes for immediate access to crypto trading

Whether this superapp comes bottom-up from Web3 wallets like Phantom, top-down from Fintech 1.0s like Revolut or Robinhood, or from entirely new players remains to be seen. All we know is the competition for control over the retail banking customer is increasingly in favor of fintechs.

Wave 2: SMB Flight

Small to mid-sized businesses (SMBs) have been famously underserved (and even remained unidentified) in part due to high servicing costs relative to returns by traditional banks, leaving room for fintechs like Revolut and Brex to swoop in.

Whereas before these fintechs relied on things like Visa APIs (e.g., Revolut’s famous multicurrency accounts), now they can plug into Bridge, BVNK, or other stablecoin orchestration APIs. The programmability of stablecoins also gives out-of-the-box treasury management capabilities like automatic cash sweeping, real-time FX, and yield optimization.

Traditional banks will have to join the race in providing better tools for their SMB client base or give up these underserved clients entirely.

Wave 3: Overall Deposit Flight

This is the existential threat to traditional banking.

The biggest advantage banks have is their cheap deposit base (they pay very little interest to depositors, often 0-0.5%, while earning 4%+ on those same deposits). Their entire business model depends on getting money for cheap and lending it out for more.

However, if depositors get access to the same end-yields banks do through stablecoins, there is real risk of deposit flight. This is why banks are lobbying against interest-bearing stablecoins it undermines their fundamental business model.

The Bottom Line

Fintech 2.0 unbundles the traditional banking value chain, redistributes the yield once captured by banks, and gives users direct access to financial primitives that were previously reserved for institutions. At the same time, it enables composable access to new markets through simple API integrations.

The question is no longer whether traditional banks will be disrupted, but how quickly they can adapt.

If you’re building in crypto or AI, feel free to reach out via Substack or LinkedIn.. I’d be happy to help you brainstorm ideas or connect you with potential investors.

Be at the forefront of the digital economy with the likes of UBS, H&M, FT, Solana Labs, Nike and many more. To receive new posts and support my work, consider becoming a free or paid subscriber.

Subscribe now

For too long, money has been synonymous with regulation, a heavily governed entity. But stablecoins usher in an era where money transforms into technology. The question isn't if this shift is happening, but how we seize this monumental opportunity. This piece dives into exactly that.

Why Are Stablecoins Even a Thing?

Why the sudden ubiquity of stablecoins?

It boils down to three powerful attributes: 24/7 instant settlement, affordability, and programmability.

While these concepts have been explored elsewhere, it's worth a quick refresher on their fundamental impact.

Instant 24/7 Settlement

Think of 24/7 instant settlement as a seismic upgrade from the traditional 24/5 banking schedule. This isn't just about convenience; it has profound implications. Primarily, it dramatically increases liquidity. Instead of funds languishing for up to seven days in intermediary accounts (like the arcane nostro/vostro systems of international transfers), money changes hands instantly. Secondly, real-time settlement significantly reduces both counterparty risk and the need for hefty settlement buffers.

What does this mean for businesses? Merchants, for example, can strategically hold idle funds in yield-bearing accounts for longer periods. Furthermore, with the reduction in settlement and credit risks afforded by instant settlement, they can allocate even more capital to these lucrative, yield-generating opportunities. In essence, businesses can retain control of their capital longer and put it to work immediately.

Stablecoins Can Be Cheaper Than Legacy Payments

Yes, stablecoins are cheaper. While there's some debate over whether this holds true given currently non-insignificant on/off-ramp fees (though these will likely reduce as more competitors join), if the money stays and settles on-chain, then this claim is certainly valid.

a16z made calculations of how much merchants save by shifting to stablecoin rails, especially in the US where the interchange isn’t capped like it is in Europe.

First, let’s go through cost savings on processing fees. In the US, credit card charges to merchants are typically 2.9% + $0.30 per transaction, whereas Stripe’s new stablecoin transaction fees are 1.5% per transaction. So, on a $5 coffee (a slightly more realistic price point than a16z’s example), Starbucks would pay $0.075 with stablecoins versus $0.445 in traditional fees, saving 83%.

In addition to saving on processing fees, merchants also save on cross-border fees. As per a16z, merchants can save 99.9% on these fees (from $44 to $0.01 on average).

Finally, merchants save by simplifying their backend reconciliation processes. For a single mid-sized merchant, annual backend reconciliation costs (labor, outsourcing, compliance) can easily range from $20,000 to $50,000 or more.

Stablecoins Create New Sources of Revenue for Merchants

Currently, under the GENIUS Act, non-financial companies (like traditional merchants) generally cannot issue stablecoins unless they receive rare, high-level regulatory clearance.

While this is still speculative, it's possible that non-bank entities like large merchants (Walmart, Amazon) will, at some point, be able to legally issue stablecoins and gain incremental passive treasury yield. After all, MercadoLibre, the Latin American e-commerce giant, has already launched its stablecoin.

Instead of large loyalty programs like Starbucks’, where customers preload funds onto their Starbucks accounts, stablecoins would allow for a similar closed-loop payment ecosystem while giving customers increased flexibility. How this would happen is that Amazon could give a 10% discount on products purchased with Amazon’s stablecoin, let’s call it “AmazonUSD.”

While skeptics might argue that merchants like Starbucks benefit from unused prepaid funds (in 2024, Starbucks generated $207.6 million in revenue from unused prepaid funds), there are a few counterarguments:

a) Consumers will be getting smarter as competitors start to offer less extractive native stablecoins that can be converted to any other digital USD version and don’t expire.

b) The addressable participant pool will increase; that is, more people will be willing to hold Amazon’s or Starbucks’ stablecoin just in case, if they know they can at any minute exchange it for USDC.

c) While merchants may lose on extractive closed-loop revenue sources like unused prepaid funds, they could gain from minting fees, and even LP fees through "LP-as-a-service" companies, or by lending out their stablecoin to market makers for extra yield.

The Stablecoin Money Flow

Recently, a bank asked me, “who gets the money” in stablecoins. The answer to this is multi-sided. Below is a simplified visualization on how the money flows and where it comes from in the stablecoin ecosystem.

Before we dig deeper into some of the key players in the ecosystem, there are interesting takeaways that come from the above visualization:

1. Stablecoins are a mechanism for packaging treasury, private credit, and DeFi yield into better consumer terms. Stablecoin treasury yield means stablecoin issuers and distributors don’t have to levy as extractive fees.

2. There are two layers that unlock additional income for stablecoin issuers and distributors: 1) yield from treasuries, and 2) efficiency gains from 24/7 settlement and real-time tracking inherent in blockchain technology.

3. Non-bank entities (e.g., large merchants) will potentially start looking more like banks in the future—think Starbucks rewards program but with real-time digital dollars instead of prefunded accounts.

4. Stablecoin distributors could become the “closed loop” payment networks (where the network controls relationships with both its cardmembers, i.e., the customers, and merchants, i.e., where cards are accepted), as opposed to current open-loop payment systems.

How Stablecoin Issuers Make Money

The short answer is that they do, and they don’t.

First, how do they make money? In three main ways: 1) money market yields, 2) minting and burning fees, and 3) listing fees (popular stablecoins get paid tens of millions of dollars by blockchains for listing on said blockchains).

While minting, and in some cases redemption, fees are “only” ~0.1%, EU and US regulation doesn’t allow the yield to be passed down directly to consumers. This means stablecoin issuers get to keep the difference. However, regulation also states that these yields have to come from the safest (and lowest yield) sources—one of the reasons why Tether has been banned in Europe. This limits their opportunity for financial engineering. These yields will likely compress in the future, not only due to lower interest rates but also increased competition.

While Tether is the most profitable company per employee ($93 million profit per employee) with 74% net margins(net income divided by revenue), Circle has below-average net margins of 9.3%, which places it under the S&P500 average of ~12%.

Why are Circle’s margins so low versus Tether? Because it pays a lot for distribution. It paid $908 million (or roughly 54% of its revenue) in 2024 alone to Coinbase for distribution.

This gives us a glimpse into how the stablecoin ecosystem might look as the market gets flooded with new stablecoins. Companies with pre-existing distribution, like Tether, will have high margins, while new upstarts, even relatively powerful ones like Circle, will have to compete (and pay) for distribution. In short, the companies that stand to gain the most here are financial networks with existing distribution, like exchanges (Robinhood, Coinbase) and banks (BofA, Citi, JPM). Imagine Binance listing fees but for stablecoins.

How Stablecoin Distributors Are the Likely Winners of the Stablecoin Wars

Where stablecoin issuance is more about regulatory arbitrage, stablecoin distribution is where most of the value will be captured. The reason Tether doesn’t have to pay $900M+ to Coinbase like Circle is that it already has great distribution, specifically among traders and in Asian markets.

I’ve created a visual categorization of stablecoin distributors below. These include: crypto apps, Layer 1s, Marketplaces, Fintechs, Banks, Payment Networks, and Merchants. While value accrual will happen at the distribution layer, there is increasing competition among them.

There are four main competitive levers that distributors can try to optimize for:

1. Retail customer wallet share: Who has the closest relationship with retail customers? Fintechs like Revolut, X, and PayPal have an edge here, as do marketplaces like Coinbase and Robinhood.

2. DeFi acceptance: Which stablecoins will be the golden standards in DeFi? The highest market cap ones could have an an edge (e.g., USDC, USDT). On the other hand, smaller, less popular stablecoins will likely pass on more of their yield via incentives, making them more competitive.

3. Decentralized trading: Which stablecoins have the deepest liquidity versus other stablecoin and crypto pairs? Most popular (highest market cap) coins have the edge here (USDC, USDT).

4. Regulatory arbitrage: Which companies can legally issue stablecoins? Circle, other first movers, and big banks like Bank of America, BBVA, and Citi have an edge here.

In summary, there are network effects: larger market cap equals more liquidity, which equals better on-chain FX rates, which then leads to higher DeFi acceptance. All this drives higher revenue, which in turn leads to better distribution (via marketing or acquisitions).

However, at this stage, it’s still unclear which stablecoins will accrue the strongest network effects.

Stablecoin Infra Will Be M&A Playground

We’ve already seen this happen through multiple investments and acquisitions. Some of them are highlighted below.

There are four trends I’m seeing related to stablecoin infrastructure:

1. Big players like Stripe and Coinbase are acquiring stablecoin infrastructure companies in efforts to become their own self-sufficient payment networks.

2. Big fintechs like Revolut or Robinhood are “co-opting” crypto innovation. We’ve consistently seen these take innovations from the crypto world and apply them to their own ecosystems, from yield accounts to tokenization.

3. Infrastructure companies are also starting from a single use case and expanding to a full-stack solution. For instance, Bridge started as a stablecoin on-ramp, but has become a full-stack stablecoin orchestration platform. Due and Mural Pay are creating payment networks, especially in emerging markets.

4. New innovative companies will come out of creative M&A deals that package different infra layers into full-stack or semi-full-stack orchestration offerings.

In sum, consolidation will occur; Stripe’s Bridge and Privy acquisitions were just the beginning.

New Financial Products Will Come to Be

While enhancing existing financial pipes is a great opportunity for incumbents and upstarts alike, I’m most excited by the net new financial opportunities stablecoins (and more broadly, tokenization) bring.

I would categorize new financial products in the following way:

1. New way of doing old things

   1. These are more efficient ways of running legacy systems—e.g., JPM Coin, tokenized private credit, instant delivery versus payment (the exchange of asset(s) and money). These are like clean tap water: they should be a basic right across the globe, but aren’t (usually due to legacy pipes). These include:

      • Not having to pay thousands of dollars just for sending money from one business to another.

      • Not having to pay 6% on average on remittances.

      • Not having to wait 3 working days to send money to a broker.

      • The ability to trade in and out of stocks whenever, wherever.

      • The ability to trade stocks at low fees.

2. New ways of doing new things

   1. This happens when new mechanisms are combined with new service models—e.g., agentic payments.

   2. Mechanism: While chargebacks have been the mechanism of keeping two transacting parties honest until now, blockchain allows new ways to maintain honesty:

      • Atomic swaps: (e.g., Agent A receives payment only if Agent B delivers service).

      • Slashing: Just like in proof-of-stake systems, a dishonest party gets their stake (e.g., in escrow) slashed and redistributed.

   3. New service/business models

      1. Investing in autonomous AI agent outputs (think owning shares in an AI agent) and AI agent-to-agent and platform payments. Most AI companies already use Stripe for payment processing (78% of the Forbes AI 50 companies build on Stripe), and while Stripe is championing stablecoin accounts as the new payment infrastructure, using stablecoins as payment vehicles for AI agents is actually not that unlikely.

      2. Creating markets for new assets via AMMs and other autonomous pricing mechanisms.

         1. Democratized payables and receivables markets—new cashflow-generating investment vehicles.

         2. “Truth” markets—utilizing financial stake for determining “objective” truth.

         3. Dynamic information markets—paying on-demand for knowledge bases used.

         4. Culture markets—prediction markets on cultural relevance, NFTs.

         5. Personal data markets—selling your own data.

         6. Public IP markets—unbundling and investing in IP (anyone can try to be the modern-day micro-Medicis).

         7. Compute markets—GPU and other granular resource markets.

These net-new asset classes aren’t just about digitizing existing value; they’re about unbundling, financializing, and creating liquid markets for intangible or previously non-tradable forms of value.

How You Can Make Money

Finally, you might say, “Okay, so what? How is this relevant to me?” It very much depends on which of these groups you belong to.

Builder

Based on recent history, the best exit opportunities seem to be in creating infrastructure that the Coinbases, Revoluts, and Stripes of the world can plug into their machinery. Incumbents are usually too slow and way behind the curve to build out services in-house. These could be wallets (Privy) or stablecoins as a service (Bridge). In the past few weeks alone, we’ve seen Coinbase acquire LiquiFi (automated payroll and token vesting), Thesis acquire Lolli (Bitcoin rewards), and Monad acquire Portal (stablecoin infra).

Particularly, bridges between the real world and the digital one are a compelling opportunity, as they are relatively new categories and ripe for innovation.

However, instead of just looking where we are currently, it would be beneficial to look where the puck is going. Ask yourself: assuming everything will run on stablecoins, what happens then?

Some first-principles conclusions are:

1. You will have to be able to swap different currencies at good rates (on-chain FX), especially low-liquidity yet popular corridors.

2. Just like with lending, at some point banks and other financial institutions will likely lobby for more lax constraints for stablecoin collateralization requirements. Risk management abilities will be key in this scenario.

3. Providing better yield for stablecoins. While stablecoins themselves cannot pass on yields, separate yield engines behind them can (yield-as-a-service), not too dissimilar from Pendle. Yield amount, regulatory compliance, and safety are key. As in DeFi, stablecoin yields will become a differentiating factor for stablecoin platforms.

4. New ways to finance short-term working capital credit (think account payables and receivables).

5. Privacy-preserving technologies (like zk) for public blockchains will become important for institutions to hide their transactions.

6. New asset types will be tokenized. Things like IP, healthcare data, transparent carbon credits, public infrastructure projects, litigations (you can fund and benefit from lawsuits), and real-time payment streams(e.g., invest in your or another specific neighborhood by purchasing residents’ recurring bill payables).

7. Embedded and programmable payments. These will be to stablecoin infra what chargebacks and other value-added services are to the incumbent card networks.

8. On-demand, machine-to-machine, and agent payments. Smart contract programmability (e.g., atomic swaps which ensure fair exchange of assets/services) as well as Coinbase’s standard x402 have a lot of potential here.

While we have some early winners already (Bridge/Stripe and BVNK), it’s still early days. If you build something better (e.g., a better stablecoin liquidity engine), the open nature of blockchains means it can be implemented by all.

The beauty of it all is that building is open to anyone. As builders solve their own problems, it will ensure a wide variety of services—a 16-year-old student from Nigeria will build something different from a 32-year-old accountant from California.

Investor

As an investor in liquid tokens (already publicly tradable tokens), the key is to find tokens that benefit from, and ideally front-run, the narratives in different token categories.

The highlighted boxes are where most current liquid investment opportunities lie.

For instance, on-chain FX sounds a lot like what Curve and Convex have been doing for years now.

For stablecoin and RWA blockchain rails, think: are there any high-throughput L1 or L2 chains that could be used to build stablecoin on?

• We saw how much SEI pumped just from being a finalist in the race for Wyoming’s stablecoin blockchain. Aptos was the other finalist. (Perhaps they read my coverage on both of them :))

  

• Arbitrum pumped from Robinhood using the chain for tokenized stocks.

• Plasma (Tether’s upcoming L1) is another obvious choice due to Tether’s distribution (however, it’s not a liquid token yet).

• Ondo launched an L1 that tokenizes institutional-grade real-world assets.

• Ethena partnered with Securitize (a leading real-world asset tokenization platform responsible for tokenizing funds for BlackRock, KKR, Hamilton Lane, and Apollo) to launch the Converge L1 blockchain.

  

On the institutional DeFi and yield side, we have protocols like:

• Maple – Institutional-grade lending with high yields

• Aave – The golden standard and is increasingly leaning toward institutional lending (white-label deal with Kraken’s Ink platform).

• Morpho – Creates new lending markets for any assets, which could cater to more exotic, long-tail assets.

• And more.

The key here is to buy these before big announcements, instead of chasing rallies.

Stocks have also become a way to invest in crypto innovation, as more and more crypto companies list there first. Ironically, the most pure-play stablecoin investment, Circle, was first available for stock market, not crypto, investors. Coinbase and Robinhood are also investable. In fact, Robinhood has been among the best-performing TradFi plays this cycle.

Kraken's ICO is also something to watch out for.

Another option (albeit my least favorite at the moment) is to invest in crypto treasury plays—i.e., companies that buy cryptocurrencies with their corporate treasury. These are essentially leveraged plays on the underlying crypto assets (BTC, ETH, SOL, HYPE). However, these have idiosyncratic risks like fund structure, which may cause investors to lose despite the underlying asset going up.

Investor (Venture Capital)

Many VCs have more ways to invest in the stablecoin boom. It’s no wonder we’ve seen a three-year high in fundraising rounds.

Some of the more interesting protocols that recently received funding:

Some of the more interesting protocols that recently received funding:

• Paradigm led the $50M Series A for Agora (white-label stablecoin service)

• On the DeFi front Kuru (a non-chain orderbook exchange protocol I mentioned in October) raised $11.6M led by

• OpenFX – FX infrastructure for real-time cross-border payments

• Limitless — Prediction markets Strategic round led by Coinbase Ventures

• Mirage – Self-repaying stablecoin loan powered by perps

• Veda – Stablecoin yield backend for institutions

• 3Jane – “Yieldcoin” backed by a pool of credit lines to crypto users and AI agents secured against future yield. Credit lines are underwritten against verifiable proofs of credit scores, onchain/CEX/bank assets, and cash flows.

• Conduit Pay – Infrastructure platform for stablecoin-based payments

• Codex - Stablecoin L1

Dragonfly has led quite a few interesting investments in the stablecoin space, making 11 investments since March 2025, from stablecoin L1s to perps on Monad.

Takeaways

While stablecoins and RWAs often get bundled together, a good mental model to think about them is this:

1. Can I trade it?(RWA, AMMs, private credit) or

2. Can I move it? (Payments).

For stablecoin companies, the key question is: will they become the new for-profit payment rails, or will they become a public good? Venture funds like Dragonfly seem to bet on the former, while a16z’s Chris Dixon believes the latter.

The value proposition for users is that stablecoins are cheaper and faster. To remain cheaper, they by nature have to have a margin lower than those of incumbents. On the other hand, payments can be the loss leaders for the ultimate cash cow, which is yield from tokenized treasuries. In this case, scale matters.

Another point is that increased stablecoins will chip away at the market share from cash. Or perhaps this will go the way of courier services, where expedited delivery fetches a premium.

Whichever the case may be, it seems that this has been the 0 to 1 moment among TradFi we’ve been waiting for years.

Opportunities abound everywhere you look.

Be at the forefront of digital assets with the likes of UBS, H&M, FT, Solana Labs, Nike and many more. To receive new posts and support my work, consider becoming a free or paid subscriber.

Subscribe now

I'm pleased to share my latest institutional cryptocurrency report, tailored for hedge funds: the Institutional Digital Asset Market Report 2020-2025.This in-depth report examines the profound transformation underway as digital assets evolve from speculative investments to strategic portfolio components for sophisticated investors.

To be honest, I haven’t seen publicly available reports like this.

Key Findings Highlight Institutional Inflection Point

My research reveals that the cryptocurrency market has reached a critical institutional inflection point characterized by:

• Regulatory clarity emerging across major jurisdictions

• Expanded market infrastructure supporting institutional-grade custody and trading

• Significant capital inflows with Bitcoin surpassing $100,000 in early 2025

• Stablecoin transaction volumes exceeding $32.4 trillion in 2024

Perhaps most telling is the dramatic increase in professional investment participation, with 47% of hedge funds now maintaining cryptocurrency exposure (up from 29% in 2023).

Four Major Investment Themes Driving Adoption

The report identifies four core themes reshaping institutional crypto strategies:

1. Stablecoins – Evolution beyond trading vehicles into essential components of the global payments ecosystem

2. Real-World Asset Tokenization – The on-chain RWA value has surged over 60% in 2024, reaching $13.7 billion with projections indicating expansion to $50 billion by the end of 2025

3. Exchange-Traded Funds (ETFs) – Bitcoin ETFs have accumulated over $40 billion in inflows since January 2024 launch, with Ethereum ETFs gathering $3.15 billion following their March 2024 approval

4. AI Agents – A nascent but potentially transformative intersection of artificial intelligence and blockchain technology

Portfolio Strategies That Outperform

My quantitative analysis presents compelling evidence for active management approaches that significantly outperform passive buy-and-hold strategies across multiple market cycles. The report details:

• Momentum-based rotation strategies with specific lookback periods delivering superior risk-adjusted returns

• Correlation dynamics between cryptocurrencies and traditional asset classes like stock indices, bonds, gold and even specific stocks

• Risk management frameworks calibrated to cryptocurrency's unique volatility profile

• Seasonal patterns and forward-looking indicators for tactical allocation

Strategic Imperatives for Institutional Investors

Based on my comprehensive analysis, I've distilled five strategic imperatives for institutional investors navigating this transformative asset class.

The full report, including detailed charts, allocation models, and case studies, is available for download through your client portal.

As always, I welcome your questions and would be pleased to discuss how these insights might apply to your specific needs via Twitter, LinkedIn or email.

Nc Analytics Institutional Digital Asset Market Report 2020 2025

11.1MB ∙ PDF file

Download

Download

Subscribe now

🔔 Quick Updates:

• Just dropped a deep dive on RTFKT with Pet Berisha - definitely check it out if you haven't already! It's a fascinating look at their evolution through the years.

💼 Career Update:

• After seven years in crypto, I'm looking to dive deeper into the VC/hedge fund side of things. If you're at (or know folks at) crypto-native funds, I'd love to connect! Reach out on Twitter or shoot me an email.

  Now, let's dive into today's analysis...

While the recent price slowdown in AI cryptocurrencies has sparked debate about the longevity of AI Agent projects, signs point to transformative innovations ahead, similar to DeFi's emergence in 2020. By now, you’ve all probably seen the chart below.

However, with new tokens launching daily, investors face a critical question: which segments of the AI Agents space offer the strongest upside potential?

Having analyzed the convergence of AI and crypto since correctly predicting AI companions and on-chain AI models in late 2022, I've observed how these early concepts are now materializing in DeFAI (AI x Decentralized Finance). The AI Wayfinder whitepaper in early 2024 first highlighted DeFAI's potential to dominate the intersection of crypto and AI. However, understanding the sector's potential requires a deeper structural analysis.

Here's a comprehensive framework for evaluating the entire AI Agent ecosystem:

The 5 AI Agent categories:

• The App

• The App Store

• The Fund

• The Framework

• Infrastructure

1. The App

We've seen multiple 10x+ gains from agents like AIXBT to most recently Fresya AI which has done a 2.5x in January alone. This category is the most customer-facing.

Valuation Cap:

1. Optimistic: Anywhere from $1B+ (Mr. Beast valuation) for entertainment agents to $22.5B (Bloomberg terminal valuation in 2023) for utility with proprietary terminals like AIXBT.

   1. Just like with the Bloomberg Terminal, the barriers to accessing the best AI Agent terminals remain high (AIXBT's terminal access costs ~$300,000). This is key to preserve exclusivity and avoid leaking too much alpha. As a consequence, for regular token holders, this means zero fundamental utility - the token serves purely as a speculation vehicle on agent adoption. Applications for the terminals could extend beyond crypto to sports betting, traditional markets, and business intelligence.

2. Realistic: $96.4M

   1. Assumptions

      1. Revenue = $5,200,000/year (Assuming Ivan on Tech's, one of the highest paid crypto influencers, pricing at 1x review per week.

      2. Advertising/Marketing Services Average EV/Revenue (as of January 2025) = 18.53

Pros:

Natural marketing pipeline — agents like AIXBT automatically market their own token through content, creating a flywheel of attention. Thus, marketing cost is effectively = zero (only pay for the cost of inference). Tokenization is also one of the few ways creators and experts can earn from their expertise in the AI landscape.

Cons:

Web2 companies like OpenAI and X could easily render some of these agents valueless via new updates. E.g. what if Grok learns to scan Twitter timeline better than AIXBT? Also, platforms like Cookie or Kaito could "democratize" access to alpha making alpha agents less necessary. AI Agents (Web2 and Web3) still have memory constraints. Higher cost of inference for more advanced models (if using APIs), high cost of training if train your own model. Cost of inference for each query – access to agent has to be limited otherwise costs could exceed revenues. Dependency on OpenAI, Anthropic APIs.

What to Look For:

Unique training data, good/unique data pipelines, developers with quantitative/trading backgrounds.

Ease of Discovery:

Medium. While AI Agents are easy to assess on the entertainment front, the alpha/investment advice is harder. For instance, some believe that AIXBT is pure alpha when in fact some of its tweets have been hallucinations if you look deeper into them. Requires backtesting and verify validity of tweets.

Revenue Model:

Fees for 1-to-1 attention.

Examples:

$AIXBT, $MOBY, $SCOUT, $DKING, $ANON, $1000x

(Not investment advice.)

Investment Score:

8/10

2. The App Store

The App Store concept is wide and is among the highest potential categories. Encompasses a vast array of platforms from AI Agent launchpads to AI gaming. Differs from AI Agents by providing a two-sided marketplace (not just protocol to user, but also user to protocol). Differs from The Framework by relying on some level of closed-source technology. Often takes longer time to build than other categories.

Valuation Cap:

• Optimistic: $280B

  • Assumptions: $30B Apple App Store 2024 revenue * Apple's 9.26EV/R ratio

• Realistic: $2.7B – $14B

  • Character.ai (an AI character “app store”) acquired by Google for $2.7B

  • Uniswap’s FDV is $14B

Pros:

Power law applies. Top platform could reach a very high market cap. Launchpads and DEXes have some of the highest revenues in crypto after stablecoins. Closed source elements create moats.

Cons:

Due to the open nature of crypto tokens can be traded anywhere. The platform with the best LP mechanism (thus lower fees) will gather liquidity. This means competitive edge may not lie in the “AI tech” but in LP and tokenomics side, making it closer to a DEX not an AI play. For instance,

The AI Agent AIXBT, with its proprietary framework, could've launched anywhere but chose Virtuals due to the attention Virtuals had as The App Store for agents. In other words AIXBT didn't utilize Virtuals “AI tech” itself due to its limited nature. Same thing occurred when the 1000xPod agent launched on Virtuals but used the Eliza framework as its tech backbone. It’s safe to say that the best AI Agents are not and will not be built with out of the box tools.

What to Look For:

UI/UX. How easily an advanced agent can be made out of the box, what is the liquidity (how good is the LP mechanism), what is the moat.

The main thing that matters for The App Store is attention. To this end the three competitive levers to pull are 1) first mover advantage (think Pump.Fun), 2) better functionality (think Eliza framework), 3) more value accrual to users via tokenomics (think Blur).

Ease of Discovery:

Medium. Easy to find market leader, but harder to understand tokenomics mean to token holders (e.g. Uniswap and its latent fee switch versus SushiSwap and its great revenue to market cap ratios, during its peak, but poor performance).

Revenue model:

Fees per agent deployment, LP fees, potential partnerships with frameworks (e.g. ability to implement Eliza for a fee).

Examples:

$VIRTUAL, Pump.Fun, DAOS.FUN, $PRIME/$PROMPT, $TOPIA, $GRIFFAIN

(Tokens with a cashtag $ indicate those projects are already tradable or will be in the near future. Not investment advice.)

Investment Score:

9/10. Invest in winners, the best in class.

3. The Fund

Similar concept and value proposition to utility AI Agents, however, biggest difference on the backend and tokenomics side. Instead of sharing alpha publicly it uses it to invest itself. Revenue share or portfolio token airdrops to holders. Future will look like investment DAO with decentralized voting on AI based strategies.

Valuation Cap:

• Optimistic: $20B – $40B

  • Using hedge fund Citadel's AUM*5% (a common valuation practice for hedge funds)

• Realistic: ~$4B+

  • The valuation of OlympusDAO at its peak in 2021

Pros:

Could be seamlessly integrated with Apps (see Section 1) for marketing purposes. Can trade around hype, and outcome not binary. Does not rely on protocols “making it”. Feasible way to democratize fundraising for talented individuals, performance over fundraising.

Cons:

Currently these funds hold a lot of their own token thus making the price movement very reflexive. Currently not yet fully autonomous, human element exists. If autonomous, risk exists that the AI capital allocator goes rogue. Potential of alpha decay (or loss of strategy profitability) with the shifting crypto metas. Faces a lot of competition if uses strictly a launchpad model.

What to Look For:

Does The Fund have systematic alpha other agents and AI fund don't. Does The Fund have interesting proprietary data or data that is hard to access by others (e.g. due to financial constraints or need for experience). Do the tokenomics benefit the holders.

Ease of Discovery:

Medium. Easy to find market leader, but harder to understand tokenomics mean to token holders (e.g. as happened with UNI).

Revenue Model:

Portfolio management fees + % of capital gains

Examples:

$VADER, $WAI, $SEKOIA

(Not investment advice.)

Investment Score:

6/10. Performance benchmarking could take a while so hard to determine good vs. bad funds. Most new funds invest in AI Agents which is a nascent market thus difficult to backtest and hard to identify luck from skill.

4. The Framework

An open-source framework that relies on user/dev flywheel.

Similar to The App Store however focus not just on launching agents but on giving the tools for users to develop more complex ones. Has many similarities with traditional crypto protocols which have open-source code thus rely on building integrations and recognition forward as the competitive edge. Relies on some sort of vertical integration to monetize e.g creator studio or an App Store.

Valuation Cap:

• Optimistic: $18B – $700B

  • $18B = Linux ecosystem value in 2023

  • $300B – $700B = Using Android’s (semi-open source) economic contribution metrics

• Realistic: $6B – $20B

  • $6B = Aave Protocol’s peak valuation in 2021

  • $20B = Chainlink’s peak valuation in 2021

Pros:

Can start out small as it could enable a flywheel where studio attracts adoption which attracts developers, which drives integration, which funds development. If framework becomes an important AI module (or Lego as they call them in DeFi) then the protocol could capture a lot of value.

Cons:

May be hard to monetize. Risks of value extractive forks that don't bring back value to the token. Lots of dependencies/ifs. "If devs adopt", "if users adopt". "Build it and they will come" doesn't always work out.

What to Look For:

Developer tech background, Github popularity (stars, forks, commits)

Ease of Discovery:

Hard. Easy to find market leader (Github popularity, market cap), but harder to know if/how the economics will take off.

Revenue Model:

Fees per use e.g. Launch fees, LP fees, usage (dev tools fees). Customer are devs.

Examples:

$AI16Z, ZEREBRO, GAME

(Not investment advice.)

Investment Score:

8/10. Hard to know if network effects have been achieved, but if they have, this could be the category with the most potential.

5. The Infrastructure Provider

Essentially a cog in the machine, the more important it is the higher valuation it could get. Could be a better data source or improved agent memory.

Valuation Cap:

• Optimistic: $22.5B – $27B

  • $22.5B = Bloomberg terminal valuation in 2023)

  • $27B = Refinitiv’s valuation in 2019

• Realistic: $135M – $20B

  • $135M= Blockworks valuation

  • $750M = Nansen Valuation Leading

  • $20B = Chainlink’s peak valuation in 2021

Pros:

Proven business model with potential for good consistent revenue. Likely targets for M&A.

Cons:

A lot of competition: Nansen, Kaito, Cookie, Dune, Flipside are all working on providing users with alpha. However, sources of alpha change constantly.

What to Look For:

User experience and usefulness of data. Quality of unique data.

Ease of Discovery:

Easy. Crypto Twitter discussions are a good indicator.

Revenue Model:

Fees to access, ad revenue

Example:

$COOKIE, Kaito AI, $ACOLYT, $GRPH

(Tokens with a cashtag $ indicate those projects are already tradable or will be in the near future. Not investment advice.)

Investment Score:

6/10. Most good projects in this category have usually been reserved for VCs but the Hype Protocol, Cookie.Fun and the Kaito AI Yaps examples could lead to more democratized opportunities.

Conclusion

• A good way to think of the AI Agent space is by replacing "DeFi Legos" with "AI Legos". An AI Agent could be launched on Virtuals, interface could be ai16z, data backend could be Acolyte, memory could be augmented by Soulgraph.

  • AI Agents will start having modular (multiple component build) vs. monolithic (all-in-one frameworks) design discussions.

• While Web3 agents have crypto-native integration advantage over their Web2 counterparts, it’s still unclear what long-term moats they may have.

• The ultimate goal for projects in the AI Agent space is to create an ecosystem of Apps, App Stores, Framework and Infrastructure.

• Value will be captured by 1) being the best in class of that category e.g. the Chainlink or Aave of AI and/or, 2) building the best vertical stack/ecosystem.

• Being close to consumer is important in this cycle (The Agent and The App Store are the closest to the consumer).

• The Frameworks could be considered as the new "Layer 1s". Just like the current Layer 1s obfuscated things like hardware and energy demands from the user, these new "Layer 1s" will obfuscate the blockchain demands from the user.

While it's hard to predict which protocols will be most valuable the same things that matter for protocol valuations today will matter in this new AI context as well:

• How well the tokenomics capture value to the holders – Fees and revenue share are good indicators

• Is there narrative/hype around it — Public/news discussion around LLMs (NVIDIA, OpenAI and DeepSeek) is a good proxy

• Protocol's replaceability in the AI hierarchy – Is the protocol the Uniswap of DEXes or the Chainlink of oracles?

• Developer tooling – How easy it is to create new advanced use-cases

Although some have said that AI Agents don't have as strong network effects as DeFi, I disagree. I believe the platform with the most use-specific training data in their segment will win. E.g. there will be no point in using the second-best AI on-chain trader.

Hence, while it's still way too early to declare winners, the power law applies especially strongly to the AI category:

More users → more training data → better outputs → more users

Given these considerations, the three most important questions you can ask yourself are:

1. Are there network effects?

2. Is there an ecosystem forming around the token?

3. Does having the token make sense (e.g. revenue share, buyback and burn, token gating, etc.)

I’ve also been in research mode analyzing what 2025 might bring for crypto markets. Instead of relying solely on my own analysis, I've gathered insights from the industry's most respected research firms.

I’m looking to build contribute more directly in the space using my seven-year experience in the space. If you are or you have connections at crypto native VC and hedge funds, please let me know via Twitter or email.

Multiple leading firms are projecting significant milestones for 2025. Bitwise forecasts Bitcoin reaching $200,000, with potential for $500,000 if government buying materializes, while Steady Lads projects a more conservative peak between $160K-220K. For Ethereum, Bitwise sees $7,000-10,000, while Steady Lads forecasts a new ATH around $5-5.5K.

However, the real story isn't in price predictions – it's in the fundamental shifts predicted by major research firms.

Three Transformative Trends

1. The AI Revolution

Bankless projects the AI agent market will grow from $10B to $250-500B, capturing 5-10% of total crypto market cap. Their analysis suggests agents will handle 30% of blockchain transactions, with individual agents processing 40-60K queries daily.

a16z's research highlights the infrastructure implications:

• AI agents operating network nodes

• Launch of the first AI-owned blockchain

• Development of TEE-secured autonomous systems

• Emergence of privacy-preserving identity solutions

2. Infrastructure Consolidation

Multiple firms point to significant infrastructure changes:

Delphi Digital and Empire podcasts predict:

• 3-5 major chains capturing majority of activity

• General purpose chains declining as specialized chains rise

• Solana, Avalanche, Celestia, Base, and "Mega ETH" emerging as leaders

Messari highlights technical developments:

• ZK technology adoption becoming exponential

• First Solana L2s launching

• Full Firedancer implementation improving client diversity

• Energy DePINs building $50-150M in supply-side infrastructure

3. Institutional Integration

Bitwise projects significant institutional developments:

• Bitcoin ETF inflows exceeding 2024's $32B

• 3% becoming new institutional allocation standard

• Department of Labor relaxing 401(k) restrictions

• Tokenized real-world assets growing from $5B to $50B

Emerging Opportunities

Gaming & Mobile

Delphi Digital predicts:

• Multiple Web3 games achieving 100K+ sustainable players

• Mobile leading Web3 gaming adoption

• High-stakes competitive gameplay driving adoption

• Blockchain-based creator payments implementation

Messari adds:

• Mobile adoption as a defining trend

• Pay-to-airdrop model replacing play-to-earn

• Higher quality mobile games launching

Payment Innovation

a16z sees merchant adoption driven by:

• Small businesses leading stablecoin adoption

• 2%+ transaction fee savings

• Coffee shops and restaurants as early adopters

• Sub-cent transaction costs enabling new use cases

Delphi Digital projects:

• Uniswap V4 launching in early 2025

• Stablecoin market cap exceeding $250B

• Wallets capturing majority of order flow value

Critical Risks & Challenges

Technical:

• Messari warns full client abstraction remains a distant goal

• DA fees may enter uncertain price discovery phase

• Increased competition from high-performance L1s

Market:

• Steady Lads predicts potential Q4 market downturn

• Trump/Powell conflict emerging as risk factor

• Post-election market uncertainty

Strategic Implications

For Builders

Infrastructure Focus:

• Prioritize specialized chains over general purpose solutions (Empire, Messari)

• Prepare for ZK technology's exponential adoption (Messari)

• Leverage AI for reducing software development costs (Haseeb)

• Consider building for mobile-first adoption (Messari, Delphi Digital)

AI Integration Strategy:

• Plan for AI agent integration with network infrastructure (a16z)

• Focus on dev tools over social agents (Haseeb)

• Prepare for decentralized AI training platforms (Haseeb)

• Consider AI-powered security tool implementation (Haseeb)

For Institutions

Market Positioning:

• Prepare for 3% becoming the new institutional allocation standard (Bitwise)

• Watch for tokenized real-world assets growth from $5B to $50B (Bitwise)

• Monitor ETF market evolution and inflows (Bitwise, Messari)

• Consider implications of bank-issued stablecoins (Delphi Digital)

Regulatory Preparation:

• Plan for cross-state legal framework adoption (a16z)

• Watch for government bond movement to blockchain (a16z)

• Monitor developments in stablecoin legislation (Bitwise)

• Prepare for new asset listing environment post-election (Messari)

For Enterprises

Infrastructure Development:

• Evaluate specialized chain solutions for specific use cases (Empire)

• Plan for integration with confidential ERC-20s (Delphi Digital)

• Consider implications of AI-owned blockchains (a16z)

• Monitor development of privacy-preserving identity solutions (a16z)

Market Opportunities:

• Assess gaming and mobile platforms for user acquisition (Delphi Digital)

• Explore blockchain-based creator payment systems (Delphi Digital)

• Consider stablecoin integration for cost savings (a16z)

• Watch for new use cases enabled by sub-cent transactions (a16z)

For Investors

Portfolio Considerations:

• Watch for Q4 market dynamics (Steady Lads)

• Monitor AI agent market growth from $10B to $250-500B (Bankless)

• Track infrastructure consolidation among 3-5 major chains (Delphi Digital)

• Assess opportunities in specialized vs. general-purpose chains (Empire)

Growth Areas:

• Watch gaming platforms achieving 100K+ sustainable players (Delphi Digital)

• Monitor energy DePIN infrastructure development (Messari)

• Track stablecoin market cap progression toward $250B (Delphi Digital)

• Evaluate opportunities in AI agent token ecosystems (Messari)

Additional Market Perspectives

While the above analysis draws from institutional research, here are some additional trends I'm watching closely:

Market Evolution

• The era of passive "buy and hold" is ending — dynamic trading becoming essential

• InfoFi (information markets based on market consensus) emerging as a new sector

• Abstract blockchain infrastructure gaining significant momentum

• NFT markets showing signs of potential resurgence

My Predictions

• AI-focused token likely to enter top 10 by market cap

• AI agent workforce development accelerating (aligned with Sam Altman's vision)

• Solana ETF approval on the horizon

• XRP will continue to demonstrate unexpected market resilience and “cultish” holder behavior

Looking Ahead

The convergence of AI integration, infrastructure maturation, and institutional adoption suggests 2025 will be transformative. However, Steady Lads' prediction of a Q4 downturn and Messari's warning about client abstraction challenges remind us to remain vigilant.

Success will likely come from understanding these intersecting trends while maintaining awareness of technical and market risks.

Analysis based on 155 predictions from leading firms including a16z, Bankless, Bitwise, Delphi Digital, Empire, Messari, and Steady Lads.

📋 Want the complete dataset? Premium members get immediate access to our detailed Google Sheet containing all 155 predictions, with firm-by-firm analysis and real-time accuracy tracking.

Take Your Analysis to the Next Level

Premium Subscription Benefits

• 📊Dynamic Tokenomics Planner Access our proprietary modeling tool for detailed token ecosystem analysis

• 📚 Full Research Archive Deep-dive reports on emerging narratives, technology breakthroughs, and market opportunities

• 📈 Real-Time Analytics Dashboard

  • Live on-chain data tracking of emerging narratives

  • NFT market intelligence and trend analysis

Read more

1) Advanced Tokenomics Planner

I've launched a comprehensive tokenomics simulation platform where projects and investors can model supply schedules by adjusting various parameters. This tool helps you visualize and optimize your tokenomics strategy with precision.

You can find the link to the site below the paywall.

2) Enhanced Analytics Dashboard

The new analytics dashboard includes:

• Web2 NFT Brand Index tracking

• ERC-20 token performance analytics by category

• Daily NFT price movement rankings

Finally, I’ve been diving deep into the AI agent landscape (research + trading), will be creating a deep dive at some point.

Coming Soon: I'm currently conducting extensive research into AI agents and will be sharing a detailed analysis in an upcoming deep dive.

Access the tools below:

Read more

Subscribe now

I. Introduction and Background

Coming out of Breakpoint 2024, Solana seems like it can do no wrong. Since its inception four years ago, there hasn't been such electric hype around the blockchain - not even during SBF's "sell me all your Solana and fuck off" phase. The numbers tell the story: fourth most valuable L1 by market cap, third biggest protocol by TVL, first in active addresses, and a casual 274x price appreciation from launch to peak. Not bad for a chain that was supposedly "dead" after FTX collapsed.

However, questions arise:

• Is blockspace just a commodity no matter which blockchain it is on or is there something special about Solana?

• Will we ever see another Solana?

There is a nonzero chance we will.

A new generation of Layer 1 blockchains believes they can challenge Solana's position. Sui, backed by ex-Meta blockchain developers, is building for gaming and TradFi applications. Aptos, sharing similar Meta DNA, takes a Web2-friendly approach to enterprise adoption. Sei focuses on high-performance DeFi. Meanwhile, pre-launch chains Monad and Berachain are redefining what a Layer 1 community can look like, combining technical innovation with memetic appeal.

This analysis explores whether any of these contenders can replicate Solana's success or if they're destined to join the graveyard of "Ethereum killers" and "Solana killers" that came before them. By examining their technical innovations, community-building approaches, and early traction metrics, we'll assess their chances of success in an increasingly competitive landscape.

II. Do Modern L1s Perform Better Than Solana?

Are They Faster Than Solana?

A new wave of Layer 1 blockchains - Monad, Sei, Aptos, Sui, and Berachain - all claim significant performance advantages. While they offer improvements in speed, composability, and development ease, each takes a distinctly different path to achieve these goals.

The three major speed categories are transactions per second or TPS (how many individual transfers or operations the network can process each second), block times (how long it takes to create a new block of transactions), and time to finality (how long until transactions become irreversible).

Performance metrics require careful analysis. Though some Layer 2 solutions boast faster block times than these new L1s, their overall time to finality remains constrained by their underlying Layer 1 chain, particularly Ethereum's finality limitations. See charts below:

The most significant advances in these new Layer 1s come from their transaction processing speed and cost structure. Some experts argue that modular blockchains (L2s) fundamentally cannot match the speeds of monolithic chains, though emerging solutions like Unichain, Arbitrum, Fuel, and MegaETH may challenge this assumption.

Whatever the case may be, Layer 1s maintain a crucial advantage: the ability to preserve common state, shared liquidity, and unified user experience - benefits that remain fragmented across Layer 2 solutions.

Are They Cheaper Than Solana?

The unique Layer 1 architecture allows these blockchains to charge substantially lower gas fees than both Ethereum and Solana. While the lower fees may not be a game changer for the average user, they become significant in two scenarios:

1. Mass adoption - when we reach billions of users

2. High-frequency trading - where performing thousands of transactions per day makes 10x lower fees critically important

Keep in mind that Solana has been congested recently due to memecoin trading activity, affecting its fee structure.

Another significant advancement in these new alternative Layer 1s is single-slot finality. Both Ethereum and Solana currently use multislot finality (though Ethereum has been looking to shift to single-slot for a while now). Single-slot finality allows transactions to finalize in a second or less, enabling:

• Instantaneous on-chain hedging

• Complex trading operations

• Reduced volatility exposure

In contrast, a multislot environment's longer finality time and potential block reorganization forces users to absorb unnecessary volatility – potentially eating into the profits of market makers and traders alike.

While the lower fees aren’t a game changer for the average user, they could make a difference for on-chain high frequency trading. Having 10x lower fees matters when performing thousands of transactions per day. Another leap forward for these new alternative Layer 1s is single-slot finality. Both Ethereum and Solana use multislot finality (Ethereum has been looking to shift to single-slot for a while now). This allows the transaction to finalize in a second or less.

How Do They Do This?

How are these L1s capable of such speeds and low transaction costs? In short, it’s because they’ve designed their blockchains from first principles.

Remember that blockchain transactions are made up four key elements:

• Execution: Processing and validating transactions

• Consensus: Achieving agreement among nodes on transaction validity

• Data Availability: Ensuring transaction data is accessible

• Settlement: Finalizing transactions and resolving disputes

The key innovations emerge primarily in execution and consensus mechanisms. In execution, two distinct approaches have emerged:

1. State Access Parallel Execution (Sui, Solana)

• Requires explicit declaration of state access (transactions have to explicitly declare which parts of the blockchain state they will access before execution)

• Enables efficient resource pre-allocation

• Increases developer complexity

2. Optimistic Parallel Execution (Sei, Aptos, Monad)

• No pre-declaration requirements

• Parallel processing without conflict checks

• Simplified developer experience

  

See the chart below for further comparison between state access and optimistic parallel execution methods:

Each chain also implements proprietary consensus mechanisms designed for speed: Sui's Mysticeti, Aptos' AptosBFT, Monad's MonadBFT, and Berachain's Proof-of-Liquidity. These mechanisms streamline block creation while maintaining security. Additionally, custom database implementations (Aptos’ Block-STM, Sui’s Objects Model, Sei’s SeiDB, and Monad’s MonadDB) optimize the data availability layer for improved retrieval speeds.

Why Speed Matters

Generally, for three reasons:

1. MEV Reduction: Faster finality significantly reduces maximal extractable value opportunities, effectively lowering transaction costs for users.

2. Enhanced User Experience: Sub-second finality enables true real-time applications:

   • Gaming: Instantaneous microtransactions

   • Trading: Precise hedging execution

   • Interactive Applications: Sub-0.1s latency (approaching Web2's 0.02s standard)

3. Institutional Appeal: Jump Crypto's investments in Sui, Sei, and Aptos validate the importance of these improvements for institutional adoption.

Performance Context:

• Visa: 24,000 TPS

• Mastercard: 5,000 TPS

• Solana (after Firedancer improvement): 600,000-1,000,000 TPS

While these L1s demonstrate architectural innovation, their performance advantages appear incremental rather than revolutionary. Block times and TPS metrics show comparable performance to Solana, and the upcoming Firedancer validator client may further extend Solana's capabilities.

However, the real opportunity lies in reliability rather than raw speed. Solana's 40% failed transaction rate, primarily from DEX-related slippage and bot activity, presents a clear vector for differentiation. For transaction-intensive applications like high-frequency trading and gaming, consistent execution may prove more valuable than marginal speed improvements.

These new Layer 1s are attempting to address these reliability challenges, provide higher speed and lower fees through their unique approaches to execution, consensus, and data availability.

III. Can Modern L1s Beat Solana?

Technical Foundations: EVM, Development, and Decentralization

As we examine whether modern L1s can challenge Solana, a fundamental question emerges: Does EVM compatibility provide a meaningful advantage? Three of the five new alternative Layer 1s (Sei, Monad and Berachain) believe it does, making EVM compatibility central to their strategy.

EVM compatibility enables seamless redeployment of Ethereum smart contracts onto these new chains. This means Solidity developers can continue using familiar tools like Foundry, Remix, and Hardhat, while users maintain their existing workflow with Metamask (or other wallets) just by changing RPC settings. While the benefits of parallelization are clear, we must examine whether EVM compatibility truly matters given Solana's success without it.

Looking at developer adoption first, Ethereum still has the lead over developer mindshare.

Not only do developers want to build on Ethereum, but it has 3x the developers (part-time + full-time) Solana has.

This advantage manifests even more dramatically in new smart contract deployment, where EVM has tens if not hundreds of thousands of contracts while Solana contracts are in the hundreds.

While the pool of builders is certainly larger for EVM chains, whether these EVM developers will benefit the new compatible L1s remains to be seen. For one thing, both Sui and Aptos (non-EVM blockchains) have 3x developers and higher developer interest than the EVM-compatible Sei has.

Beyond developer metrics, TVL presents a more complex picture. While EVM chains have a clear lead over non-EVM chains (Ethereum alone has over 50% of the TVL at the time of writing), the EVM-compatible chains Aptos, Sei, Sui and Monad won't get a portion of that TVL automatically. Analysis of Ethereum Layer 2s suggests a more modest potential: around 2% TVL unlock for EVM Layer 1s like Berachain and Monad, as these could be perceived as more similar to Ethereum, something some users may value.

The decentralization question adds another layer of complexity. While lower hardware requirements are often associated with Ethereum and to some extent with the EVM ecosystem, the reality is more nuanced. Ethereum's total validator count far exceeds Solana's, but these validators are concentrated in the Lido protocol. What matters more for decentralization are the block-producing validators.

Solana takes a different approach. While its hardware requirements are higher, its vision emphasizes democratized verification rather than validation, providing transparency over the blockchain. Notably, Solana has no minimum staking requirements, unlike Ethereum's 32 ETH requirement for solo validators.

To understand the theoretical potential for decentralization across these new L1s (i.e., can anyone with a high-end MacBook Pro run a validator node), I analyzed the minimum specs for running a node on each blockchain. Solana shows the most prohibitive specs while Ethereum has the least. Perhaps unsurprisingly, the new non-EVM Layer 1s Sui and Aptos align closer to Solana's higher-end hardware requirements.

Yet in practice, we see little correlation between validator count (or decentralization) and EVM compatibility at the time of writing.

The lower validator counts on these new L1s partly reflect their higher hardware requirements. While we must consider their youth and commitment to making blockchain verification (rather than validation) accessible, none can yet claim to surpass Solana in decentralization.

These higher hardware requirements result in lower validator counts across the new L1s. Though they're still young and focus on accessible blockchain verification rather than validation, none currently rival Solana's (much less Ethereum’s) decentralization metrics.

The Case Against EVM Compatibility

Modern cross-chain solutions have transformed the user experience landscape. With DeBridge, LayerZero, and Wormhole offering seamless and cost-effective bridging, the distinction between EVM and non-EVM chains has blurred from a user perspective.

Instead, chains now compete on three key fronts:

1. Transaction fees: E.g. Aave on Arbitrum is cheaper post-fees than Aave on Ethereum. However, with most new L1s offering similarly low fees, this is quickly becoming a commoditized feature.

2. Security (including decentralization): Critical for institutions and large stakers. Ethereum currently leads here, benefiting from the Lindy Effect, while these new Layer 1s face initial disadvantages in gaining trust.

3. User-valued applications: Perhaps most crucially, users follow compelling use cases regardless of chain loyalty. Even the most dedicated Ethereum supporters bridged to Solana for NFTs, just as they now explore Solana's thriving DePIN ecosystem. Use cases trump allegiances.

The data reinforces this user-centric view. Non-EVM chains show more than 3x more active wallets according to a16z (though some of this reflects bot activity).

Moreover, Solana's ecosystem shows stronger cross-chain appeal, with more inflows from Ethereum than outflows back to it.

These patterns suggest that EVM compatibility alone doesn't guarantee network effects or liquidity – as evidenced by numerous dead EVM-compatible chains. Some of them seen below.

Base's success illustrates this point: its growth stems more from Coinbase's established user base than from EVM compatibility.

In summary, while EVM compatibility may not completely eliminate user experience friction, Ethereum still captures the majority of developer mindshare and offers significant advantages. The dApps built by these developers can be ported over to EVM-compatible chains, though this doesn't guarantee user adoption. Ethereum also has promising developments in ZK technology, including account abstraction (Safe adoption, ZKEmail, EIP-7702) and ZK-SNARK applications (Zupass, Railway, 0xbow, Rarimo). The analysis shows decentralization depends more on hardware requirements and design choices than EVM compatibility, and TVL doesn't automatically transfer to new EVM chains. Most importantly, user adoption follows compelling applications rather than technical compatibility, explaining why non-EVM chains like Solana can thrive. This suggests new L1s need more than EVM compatibility to compete effectively.

Next, we will look deeper into the competitive advantage of each of these new Layer 1 blockchains.

User-Centric Gaming & TradFi Innovation – Case: Sui

Founded by the ex-Meta blockchain team and launched in May 2023, Sui has emerged as a prominent Layer 1 contender. While Mysten Labs' CEO Evan Cheng positions Sui as a general-purpose blockchain, its architecture reveals a natural advantage in gaming and social applications, driven by three key strengths: seamless user experience, powerful gaming primitives, and TradFi-ready infrastructure.

User Experience & Gaming Features

First, seamless user onboarding removes traditional Web3 friction points. Users can create wallets using Gmail accounts and Face ID, while gasless transactions at the protocol level eliminate the complexity of managing gas fees. This foundation enables mainstream adoption of blockchain gaming.

Second, Sui's technical architecture directly enables new gaming experiences:

• Dynamic NFTs evolve with player achievements. For example, in Legends of Arcadia, character NFTs can upgrade based on player decisions and in-game accomplishments

• Zero-lag gameplay through efficient transaction processing enables real-time multiplayer interactions

• Asset composability (similar to ERC6551) where objects can own other objects

• Programmable Transaction Blocks (PTB) handling up to 1,024 operations at once, enabling complex in-game actions like:

  • Mass crafting of items

  • Trading entire inventories in one transaction

  • Executing complex multi-step quests

• The Sui Kiosk system powers new economic models:

  • Players can run their own in-game fashion stores

  • Artists can create in-game galleries to sell generative art NFTs

  • Sui primitive also allows NFTs to be sold only if conditions are met. These conditions could dictate that 1) royalties are paid and 2) they are sold only via a specified marketplace.

• Sui’s Time Capsule NFT primitive allows NFTs to be revealed at a specific time giving rise to:

  • Timed weapon or character reveals

  • Scheduled in-game events

  • Raffle reveals or in-game betting

• Obfuscated NFTs create novel monetization options:

  • Try-before-you-buy item previews

  • Progressive content unlocks

  • Premium feature reveals

These capabilities are further strengthened by robust infrastructure:

• SCION (Scalability, Control, and Isolation on Next-Generation Networks) integration providing DDoS protection and improved connectivity

• Mysten Labs-built Walrus protocol enhances Sui's scalability by moving away from traditional Web2 storage solutions

• SuiPlay0X1, a dedicated gaming console with native support for Sui Web3 games (Solana is responding with its own gaming console Play Solana)

• Sui’s Web3 gaming discovery and education website

• Company-specific loyalty points using the Closed-Loop Token standard

Sui also emphasizes Web2 company adoption through several key features:

• The Closed-Loop Token standard enabling company-specific loyalty points and rewards programs

• Enoki, a SaaS platform from Mysten Labs, allowing Web2 companies to seamlessly integrate with Sui for:

  • In-game asset trading

  • Loyalty program management (I’ve written about their benefits extensively here)

  • Custom token implementations

    

Developer support is also particularly strong:

• 1:1 engineering office hours with the core team

• Move programming language designed to minimize common smart contract vulnerabilities like re-entrancy and arithmetic overflows (in 2023 almost $2B of value was lost due to hacking)

• Comprehensive documentation and development resources

  

Beyond gaming, Sui targets institutional adoption through two key offerings:

First, its Closed-Loop Token standard (i.e. permissioned tokens and the equivalent of Token Extensions on Solana) supports:

• Regulated digital currencies

• Corporate tokens and rewards

• Restricted investment products

• Time-locked compensation

• Regulatory sandboxes

• Regulatory-compliant trading

• Programmable grants or subsidies

• Tokenized carbon credits

Second, DeepBook, Sui's central limit order book, brings institutional-grade DeFi capabilities:

• High-frequency trading applications leveraging Sui's fast finality

• Aggregated liquidity pools supporting large-scale trades

• Advanced trading strategies including limit orders and flash loans

• Professional trading features that move DeFi closer to traditional finance capabilities

Current Ecosystem Status

Sui’s early ecosystem traction shows promise across multiple sectors. It’s most used dApps by sector are:

• Social: RECRD – A Web3 TikTok competitor with 481K daily active addresses

• Gaming: Birds – 163K daily active addresses. A Telegram mini app so take it for what it’s worth).

• DeFi: NAVI Protocol – $294M TVL – The Aave of Sui, but still only 3% of Aave’s TVL

Sui's 2024 Overflow Hackathon showcased both ecosystem potential and challenges, drawing 352 project submissions from 79 countries. The winning projects across different categories revealed interesting insights about Sui's competitive advantages:

In gaming, where Sui's unique features shine, AresRPG won first place with its procedurally generated MMORPG world – a project that leverages Sui's gaming-specific capabilities.

However, winners in other categories highlighted the platform's current limitations:

• Consumer category winner Pandora Finance (a decentralized prediction market for token prices, events, and commodities) has only 2.5K users and doesn't justify Sui's high-performance features. Without millions of concurrent users or an on-chain order book, such platforms could operate effectively on simpler chains like Polygon, as Polymarket has shown.

• DeFi winner Hop Aggregator (Swap aggregator meets Pump.fun), while valuable for the ecosystem, largely replicates functionality already available on Solana

The analysis of Sui reveals a clear pattern: while it offers comprehensive Layer 1 capabilities, its true competitive advantage against Solana lies in specific verticals rather than general-purpose functionality. Its gaming-focused features, from dynamic NFTs to sophisticated in-game economies, represent meaningful innovation rather than mere technical iteration. The platform's early traction (with 481K daily active addresses on RECRD and significant gaming engagement) suggests this focused approach can work.

However, Sui's experience also demonstrates a broader truth about competing with Solana: technical capabilities alone aren't enough. While Sui offers impressive features like DeepBook for DeFi and Closed-Loop Tokens for institutions, the hackathon results show that many applications don't yet require such high performance. Success appears to come from identifying and dominating specific verticals (in Sui's case, gaming and Web2 integration) rather than trying to outperform Solana across all dimensions.

As we examine other Layer 1 competitors, this lesson about focused differentiation versus general-purpose competition will prove crucial.

High-Performance DeFi Infrastructure – Case: Sei

Founded by an ex-Robinhood engineer and Goldman Sachs investment banker, Sei launched in August 2023 with an initial focus on DeFi. While now positioning itself as a general-purpose blockchain, its technical architecture reveals its DeFi roots.

Technical Architecture & Features

Sei offers EVM compatibility with a unique twist: while Ethereum and Sei wallets have different public addresses, they share the same private key, enabling cross-chain access.

However, this compatibility still requires users to complete a five-step process, somewhat limiting its practical benefit.

Like other new L1s, Sei offers low gas fees through parallelization. One of Sei's points of differentiation is its proactive approach to MEV. It uses Frequent Batch Auctioning (FBA) to prevent MEV frontrunning, thus making the transaction fees for users lower, and a large portion of MEV value is captured and redistributed to SEI token holders, validators, and delegators.

While currently showing the lowest block-producing validator count, Sei aims to improve decentralization through zero-knowledge light clients, enabling broader blockchain verification.

Current Ecosystem Status

Despite launching before Sui, Sei's ecosystem shows more limited adoption:

• 52K active addresses (versus Sui's 1.1M)

• $149M TVL (approximately 15% of Sui's)

• Ecosystem dominated by standard DEXs and money markets

Some of the more interesting apps include DragonSwap, a DEX. DragonSwap where stakers share in the revenue through a fee switch and receive airdrops. The DEX also is looking to integrate a predict-and-earn platform.

Another interesting category of development is perpetual DEXs like Filament and Vertex. Sei’s speed enables perpetual DEXs. However, the competition in this product category is increasing with dApps like Sui’s DeepBook and HyperLiquid.

With Sei, the intrigue lies not in its current ecosystem, but in its ambitious vision for the future.

With Sei, the real interest lies not in its current ecosystem but in its future plans.

Future Development Initiatives

Sei is actively working to expand its ecosystem through several funding programs:

Sei Creator Fund (with Gitcoin):

• $10M fund launched June 2024

• 646 applications (exceeding Sui's Overflow Hackathon)

• Categories of focus include:

  • Creative media and IP development (e.g. lore around Sei)

  • Decentralized content platforms

  • NFT marketplaces and tools

  • Social Platforms and fan engagement

  • Decentralized crowdfunding

  • Web3 creative tools and infrastructure

  • Creator DAOs and collectives

  • Web3 content discovery, creation and curation

  • Web3 skill-sharing and education