Tom Lee vs Jordi Visser: The AI Trade's Next Leg Runs Through Ethereum — If Agent Rails Can Be Built in Time

Two macro analysts looked at the same AI trade this week and saw different endpoints. Tom Lee, Fundstrat's co-founder, sees a second leg. Jordi Visser, who runs AI research at 22V Research, sees the easy money already banked. Yet both named the same settlement ledger. That convergence deserves more attention than the disagreement. The split is partly a matter of historical framing. Lee covered mobile phones as an analyst in the early 1990s. Motorola and the infrastructure suppliers led that cycle's first leg. The larger winners arrived later, usually as tower companies spun out of carriers, and then Apple. Lee expects the same shape for AI, with financial services as the downstream market. Visser spent two decades at Weiss Multi-Strategy Advisers, latterly as chief investment officer, and he now expects roughly 30% annual returns from AI rather than the seven or eight times early investors chased. Lee reads that compression as rotation. Visser reads it as maturity. History doesn't repeat, but the mobile analogy is suspiciously neat. The substantive disagreement runs through a simple question: can software agents use the banking system? Lee listed trust, proof of funds, lending, and tax collection as the four reasons human commerce built itself around banks. Agents need none of those, he argued. Bank ledgers settle in a single national currency. Agents settle in whatever the counterparty accepts. Money, in Lee's telling, is becoming code. Equities, gold, and tokens could all clear as payment if the rail recognizes them. "It's a mistake to think that this is going to be built on traditional financial rails," he said. The most concrete piece of that alternative rail is ERC-8183, a proposed Ethereum standard filed on Feb. 25. It locks an agent's payment in escrow until a designated evaluator signs off. Ethereum Foundation researcher Davide Crapis co-authored it with three Virtuals Protocol engineers. It carries Draft status, so nothing about it is final. That last sentence is doing more work than it looks. An escrow contract is older than banking, but splitting evaluation from settlement into a machine-readable primitive is genuinely new. The evaluator becomes a new trusted node, and that node is a fee collector. In my audit experience, the first thing I look for in escrow designs is who holds the keys during the evaluation window. ERC-8183 does not answer that question yet because it is not final. That uncertainty is the real story. Standards are not technical artifacts; they are political settlements. A standard that lets agents transact but leaves evaluator selection ambiguous will simply move the trust problem from the bank to the marketplace. Liquidity is the only truth in a world of noise, and liquidity will flow toward whatever standard reduces that ambiguity first. Now the uncomfortable part of Lee's thesis. He chairs BitMine Immersion Technologies, the largest corporate holder of ether. The company disclosed 5.79 million ETH on July 27, close to 4.8% of circulating supply. Crypto and cash holdings reached $11.8 billion. BitMine states the dependency plainly in its own investor materials. "So our future price for Bitmine stock is heavily dependent on the future price of Ethereum," Lee said in the July chairman's message. He put the correlation between BitMine shares and ether at 90%. Anyone weighing his agent thesis is also weighing that balance sheet. A straightforward critic would call this a conflict of interest and move on. That is too easy. The conflict cuts both ways. His balance sheet makes Ethereum price an input, not an output. When Lee says the next AI leg runs through crypto payment rails, he is describing the world his treasury needs. That does not make him wrong. It makes his timeline untrustworthy. If agent commerce is genuinely five years away, BitMine is sitting on an $11.8 billion treasury with a 90% correlation to an asset that has fallen 51% over twelve months. Visser's 30% annual return target is not bearish. It is a demand that AI start producing revenue instead of valuation. The numbers on the agent-economy side do not support a near-term leap. Jansen Teng, co-founder and chief executive of Virtuals Protocol, shared the panel with Lee. His platform lets agents hold wallets and pay each other onchain. Teng said the launchpad for agent tokens has cleared about $15 billion in trading volume. Agent-to-agent commerce has settled roughly $500 million in a year. Speculation on agents is therefore some thirty times larger than agents transacting. Both figures are company-reported and have not been independently verified. Teng also said the agents kept $2.5 million in profit, and that the product has not reached product-market fit. Virtuals commissioned the Fundstrat research and is a client of the firm. A $500 million settlement pool is not nothing. It is also not a second leg. It is a seed. The mobile cycle's second leg produced Apple, a company with actual earnings. The agent cycle's second leg has produced a $2.5 million profit pool and a token 89% below its January 2025 peak. The gap between $15 billion of speculation and $500 million of settlement is the gap between narrative and utility. Value is the illusion we agree to sustain, and right now the market is sustaining an illusion at thirty times the rate of use. The contrarian angle is not that Lee is wrong because he is biased. It is that both Lee and Visser may be right in sequence. Lee is right that traditional rails were built for human identity, not machine identity. Visser is right that the easy money has been made on compute. The next phase depends on a standard that may never ship in its current form. ERC-8183 is Draft status. The evaluators it depends on do not yet have a reputation layer. The agents holding wallets have no credit history and no domicile. Building a financial rail for them means rebuilding the trust stack that banks spent three centuries constructing. So the question is not whether the AI trade ended. It is whether machine payments arrive before the balance sheets betting on them need the story to work. Watch the ratio, not the price. If agent-to-agent settlement starts moving from $500 million toward $15 billion, the narrative will have found its tower. If it stays flat, the 30:1 speculation-to-commerce ratio will correct, and that correction will look like a macro crash rather than a narrative reset. That is not a forecast. It is just liquidity being honest. Chaos is just liquidity waiting for a narrative. But a narrative without settlement is only a prayer.