
Tokenization has been a crypto promise for years, and Arthur Hayes says that history has already shaped market expectations.
What could change the trade, in his view, is AI: agents managing portfolios, collateral and payments could turn tokenized assets into part of a constantly operating on-chain financial system.
Key Takeaways
- Hayes sees AI agents as a potential new source of demand for on-chain financial infrastructure.
- He is looking beyond new launches to established protocols with users, liquidity and revenue.
- Tokenization can increase activity on blockchains without automatically increasing a token’s value.
The question is which protocols will handle the trading, settlement and liquidity if tokenized funds, bonds and other assets bring more capital on-chain, and whether their tokens receive any of the resulting economics.
Hayes’s tokenization case begins with AI agents
Hayes sees tokenized stocks or Treasuries as raw material for an automated financial system. “AI agents will run all of your financial services,” he said, arguing that cheap compute could enable software to rebalance portfolios, route payments and trade around the clock.
Such a system needs rails capable of handling many small and frequent actions: liquid markets, reliable settlement, collateral management and clear payment permissions. The practical problem is already visible in early agent-payment tooling. XRPL’s AI payment tools, for example, separate an agent’s ability to prepare a transaction from the authority to sign it. Scaling agentic finance means solving that operational layer as well as creating more tokenized assets.
How Hayes connects tokenization to a token re-rating
Tokenized assets
Funds, bonds and credit move on-chain
→
AI agents
Automated users need 24/7 financial rails
→
Protocol activity
Settlement, liquidity and fee generation
→
Token re-rating
Possible only if value reaches holders
The sequence reflects Hayes’s thesis. A rise in tokenized assets alone does not complete the final step.
Large RWA totals do not settle the liquidity question
Tokenization has already produced large headline figures, though the figures need to be read carefully. On October 12, RWA.xyz listed XRP Ledger with $5.05 billion in tokenized assets excluding stablecoins. The same dashboard marked 9.8% as distributed, showing that an asset total, broad ownership and active secondary-market liquidity are not interchangeable measures.
XRPL ranking found the same gap between the network’s total value and the portion RWA.xyz labels as distributed. Issuer restrictions, investor eligibility and redemption terms can all shape how useful a tokenized product is, even when its underlying value is recorded on a public blockchain.
Hayes applies a more selective filter. He said he looks for what is “most misunderstood that could re-rate,” rather than assuming every company or token associated with real-world assets will benefit equally. A protocol needs a role in the actual financial workflow before tokenization can become meaningful demand for its token.
Existing revenue gives the thesis something to measure
The current data show the type of on-chain businesses Hayes has in mind. At the time of writing, DefiLlama listed Hyperliquid with $7.03 billion in total value locked and $71.05 million in protocol revenue over the previous 30 days. The figures do not make Hyperliquid a direct tokenization play, but they show a protocol with measurable use, income and token-holder economics before any further institutional adoption arrives.
Pendle offers a different example. The yield protocol had $1.22 billion in TVL, $419,752 in 30-day protocol revenue and $785.66 million in 30-day DEX volume at the time of checking. Its tracked revenue model assigns 80% of revenue to sPENDLE holders. Activity, fees and a stated distribution mechanism can therefore be assessed separately from a market narrative.
| Protocol snapshot | Current operating data | Why it fits Hayes’s screen |
|---|---|---|
| Hyperliquid | $7.03B TVL; $71.05M 30-day protocol revenue | A live business with measurable activity and revenue. |
| Pendle | $1.22B TVL; $419,752 30-day revenue; $785.66M 30-day volume | Revenue distribution gives token economics a testable component. |
Figures from DefiLlama at the time of writing. They are illustrations of the framework, not token recommendations.
Why Hayes is looking beyond new launches
Hayes believes agents are more likely to rely on infrastructure that has already survived volatile markets. He pointed to the protocols that remained active after DeFi Summer, with existing liquidity and users, rather than a new wave of low-float tokens facing large unlocks. “The stuff that’s going to 10x is going to be already tradable right now,” he said.
Operating history does not eliminate risk. Smart-contract failures, competition, regulatory restrictions and token dilution can still break the link between a growing protocol and its token. Hayes’s framework starts with a tougher requirement than a new announcement: the protocol must already be useful, and the token’s claim on the business must be clear enough to examine.
Tokenization expands the market; value capture decides the winners
Hayes’s thesis rests on a narrower claim than “tokenization lifts crypto.” Tokenized assets may bring more capital on-chain, while AI agents may create more frequent demand for financial infrastructure. Protocols that already process activity and can return part of the economics to holders would have the clearest route to a re-rating if that activity grows.
“Price is what you pay, value is what you get,” Hayes said. Tokenization may expand the addressable market, but it does not answer who collects the fees, who owns the liquidity or whether a token has a claim on either. Those questions may decide whether the theme becomes a broad crypto bull run or another story that leaves token holders behind.
Disclaimer: This article reports Arthur Hayes’s views and provides general market analysis. It is not investment advice.



Be the first to comment