What I Told Congress About the US-China AI Race

Key Takeaways

• U.S. models still outperform Chinese counterparts on most benchmarks
• China accelerates AI diffusion through cheaper, open‑source models
• Export controls slowed Chinese compute but boosted its domestic semiconductor push
• Energy capacity is China’s advantage; U.S. faces grid cost pressures
• Congress must align AI policy with job security and affordable energy

Pulse Analysis

The U.S. House Select Committee on China recently heard testimony from an AI expert who framed the America‑China rivalry as a multi‑dimensional race. While acknowledging that American firms dominate the frontier model leaderboard, the witness warned that victory cannot be measured by raw performance alone. He highlighted China’s strategic focus on rapid adoption, especially in physical AI such as robotics and autonomous vehicles, and stressed that the competition now extends beyond algorithms to infrastructure, supply chains, and societal impact.

On the technical front, the testimony underscored three pillars of AI supremacy: model quality, compute stack, and energy availability. The United States retains a decisive edge in model performance, backed by deep capital markets and a talent pipeline that attracts global researchers. Conversely, China has turned energy abundance into a competitive lever, building power capacity equivalent to the entire U.S. grid in four years, and is racing to close the semiconductor gap by domesticating photoresists, lithography tools, and advanced packaging. These moves narrow the compute advantage the U.S. once enjoyed.

The policy implications are clear: export controls have bought the United States a brief window, but they also spurred China’s home‑grown chip ecosystem. To convert technical lead into broad‑based prosperity, lawmakers must craft AI legislation that tackles job displacement, energy cost inflation, and equitable access to AI tools. Aligning federal research funding with domestic manufacturing, incentivizing low‑carbon data centers, and establishing workforce retraining programs could ensure that AI fuels productivity without widening inequality. In the witness’s view, Congress holds the decisive lever to turn AI competition into a win for ordinary Americans.