Raoul Pal: U.S.-China AI Race Has No Clear Winner

Raoul Pal, co-founder of Real Vision and a former Goldman Sachs hedge fund manager, said May 18 the contest between the United States and China over artificial intelligence targets the systems that generate intelligence rather than territory or weapons. He spoke at Consensus 2026 in Miami and posted remarks on X.

Pal wrote, “The U.S.-China AI race is a race no one can win and no one can afford to lose. Every great power competition in history was for territory, resources, or weapons. This one is the first that is for none of them. It is a race for the substrate of intelligence itself.”

At the conference Pal proposed a plan he called “Universal Basic Equity.” Under the proposal, citizens would receive ownership stakes in AI systems to share in the value those systems produce as AI automates knowledge work. He said ownership models built on cryptographic tokens and blockchain-style distribution could be used to allocate those stakes.

Pal argued the proposal is a response to expected large-scale automation of white-collar and knowledge jobs. He said broad ownership in AI systems could provide an alternative to relying only on wages or government transfer programs as a source of income for people affected by job displacement.

Pal also described technical differences between the U.S. and China. He noted the United States maintains advantages in high-end compute capacity, top-line model performance and frontier large language model work. He cited a May 2026 analysis finding China gaining in areas such as model efficiency, rapid domestic deployment, integration of AI into manufacturing and the ability to train competitive models with less compute than U.S. frontier labs typically use.

Those technical differences reflect different strategies: U.S. efforts have focused on scaling compute and pushing leading model capabilities, while China has emphasized efficiency gains, open-source sharing of models and embedding AI in industrial and physical systems. Analysts tracking the competition say the outcome depends on multiple measures, including compute, model quality, adoption, integration and deployment.

Pal noted that existing trade and tech tensions affect the environment for both AI and digital assets by shaping export controls, chip access and cross-border technology flows. He said cryptographic ownership and token-based distribution operate across jurisdictions and can be affected by those rules.