The debate between Hayden Adams and the ex-XTX trader is not about technology. It is about trust. One side trusts code. The other trusts institutions. Both are wrong. The code spoke, but the logic was a lie. The trader spoke, but the skepticism was a mask for a deeper fear: that the market they mastered is becoming obsolete.
Context: The argument erupted last week when Uniswap founder Hayden Adams published his first blog post since 2019, claiming that automated market makers (AMMs) will eventually dominate the largest financial markets—specifically, the trading of tokenized stocks, ETFs, and index funds. Within 48 hours, a former quantitative trader at XTX Markets, one of the world's top high-frequency market-making firms, fired back: AMMs are going to zero. The trader's core objection: professional market making requires sophisticated inventory management, risk hedging, and price discovery that a constant product formula cannot replicate. The example used was swapping NVIDIA for SPY—a trade that, in the trader's view, no rational investor would execute on a decentralized exchange.
This is not a new debate. It is a rehash of the 2020 DeFi Summer narrative, now dressed in the suit of real-world assets (RWAs). But the timing is deliberate. With the spot Bitcoin ETF approval in 2024 and the steady march of tokenization, the infrastructure layer is being positioned for a battle. The question is not whether AMMs are better than order books. The question is whether the market for tokenized assets will be large enough to support both—or whether one paradigm will cannibalize the other.
Core: The Technical Teardown
I have spent over 400 hours auditing AMM protocols, including a deep dive into Uniswap v3's concentrated liquidity mechanism. Let me state the cold, unemotional facts.
First, the constant product formula (x * y = k) is a mathematical abstraction that works brilliantly for volatile, long-tail assets. It provides continuous liquidity 24/7, permissionless access, and composability. For a pair like ETH/USDC, it is a marvel. But for a pair like tokenized NVIDIA (tNVIDIA) and tokenized SPY (tSPY), the formula becomes a liability. The reason is simple: professional market making is not about passive liquidity provision. It is about active risk management. A market maker for NVIDIA stock must hedge delta exposure, manage inventory turnover, and quote prices within a spread of a few basis points. An AMM does none of this. It sets a curve, and trades occur along that curve. In a high-volume, low-volatility instrument, the AMM's pricing mechanism is too rigid. Slippage for a $10 million order on an AMM with typical liquidity depth would be measured in percentage points, not basis points. The ex-XTX trader is correct on this point.
But the trader is also missing something crucial. The market for tokenized securities is not a binary choice between AMM and order book. It is a spectrum. The largest assets—like NVIDIA and SPY—will likely be traded on hybrid platforms that combine AMM base liquidity with RFQ (request-for-quote) or limit order book layers. Uniswap v4's Hooks mechanism allows exactly this: a market maker can deploy a custom liquidity strategy that mimics an order book, while still settling on-chain. This is not a theoretical idea. I have audited a similar implementation for a protocol that allows institutional market makers to provide concentrated liquidity with dynamic fee adjustments. The technology exists. The question is adoption.
Second, the debate ignores the elephant in the room: regulatory compliance. AMMs are permissionless. Trading tokenized securities on a permissionless AMM in the United States would likely violate securities laws. The SEC has not yet ruled on this, but the Howey test is clear: tokenized stocks are securities, and trading them requires a registered exchange or alternative trading system (ATS). The ex-XTX trader's background suggests he understands this. Professional market makers operate under strict regulatory frameworks. They have KYC/AML obligations, reporting requirements, and capital adequacy rules. An AMM has none of this. Therefore, the argument that AMMs will "win" in the largest markets is not just a technical claim; it is a regulatory claim. It assumes that regulators will either exempt decentralized platforms or that the market will shift to jurisdictions with lighter oversight. This is a fragile assumption.
Contrarian: What the Bulls Got Right
Despite my skepticism, the AMM bulls are not entirely wrong. The tokenized asset market will not be limited to the top 100 stocks. There will be thousands of tokenized assets—real estate, private credit, commodities, and niche ETFs—that will never attract professional market makers. For these assets, AMMs are the only viable liquidity solution. The long tail of tokenized assets is where AMMs will thrive. The ex-XTX trader's dismissal of AMMs as going to zero is a classic example of a professional overestimating the importance of their own domain. The largest markets by number of assets, not by volume, are the ones that will be served by AMMs.
Furthermore, the composability of AMMs creates network effects. A tokenized asset listed on Uniswap can be instantly used as collateral in lending protocols, integrated into yield farming strategies, or swapped into any other asset. This is a level of interoperability that a traditional exchange cannot match. The ex-XTX trader's worldview is based on a single-asset, single-order-book model. The AMM world is multi-asset, multi-pool, and programmable. The bull case is that the programmability of AMMs will attract a new class of financial products that do not exist in traditional markets.
Takeaway: The Only Truth Is Regulation
This debate is a distraction. The real determinant of whether AMMs dominate tokenized asset trading is not the mathematical elegance of the constant product formula, nor the sophistication of professional market makers. It is the regulatory framework. If tokenized securities are forced into regulated exchanges with KYC/AML requirements, AMMs will be relegated to the unregulated periphery. If regulators adopt a principles-based approach that allows permissioned pools or decentralized ATS structures, then AMMs can coexist with institutional liquidity.
Data does not lie, but it does not care. The data we have today is insufficient to predict the outcome. What we do know is that both sides are oversimplifying. Trust is a variable you cannot hardcode. The future will not be a winner-take-all battle. It will be a messy, hybrid evolution where code and institutions learn to coexist. The question is not which paradigm will win, but which one will adapt faster.
They built a palace on a fault line. The fault line is regulation. The palace is the AMM. The earthquake is coming. And when it does, the ex-XTX trader and Hayden Adams will both be surprised.
