Fundstrat's Tom Lee argues ETH is the infrastructure layer for an AI economy that needs neutral settlement rails
While everyone was busy chasing the next Nvidia, Ethereum quietly ate AI hardware’s lunch. Over the past month, ETH outperformed the Roundhill Memory ETF (DRAM), a basket of global memory chip companies tied to the AI boom, by 55 percentage points, according to Fundstrat’s Tom Lee.
The case for ETH as AI infrastructure #
Tom Lee laid out his thesis around July 17, framing Ethereum not as a speculative crypto bet but as critical infrastructure for the AI economy. His argument hinges on a simple idea: as autonomous AI agents proliferate, they need a neutral, programmable system to move money around. Lee says Ethereum provides exactly that, acting as the “guardrails” of consumer trust in a world where machines increasingly transact with other machines.
The Roundhill Memory ETF launched on April 2 and focuses on global memory chip companies, the kind of firms that supply the physical hardware powering AI workloads.
Institutional money is noticing #
Significant inflows have been reported into spot ETH ETFs, with BlackRock’s ETHA drawing particular attention.
Both KuCoin and Binance have recently amplified this narrative through their research and media channels, reinforcing the idea that Ethereum occupies a unique position in the AI-related financial ecosystem.
What this means for investors #
Lee’s core argument represents a strategic pivot: instead of betting purely on which chipmaker wins the AI hardware race, investors could gain exposure to the transactional layer that all AI agents will eventually need, regardless of which hardware they run on.
There is also competitive risk. Ethereum isn’t the only blockchain vying to become the settlement layer for machine-to-machine transactions. Solana, various Layer 2 networks, and even purpose-built AI chains are all competing for the same role.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our