IFP report: US leads China in AI by only ~6 months, existing policy tools could double it

US think tank Institute for Progress (IFP) released a 23,000-word research report on the US-China AI competition on October 20. The study took six months and was authored by Georgia C. Adamson, Konstantin Pilz, and Jeremie Harris. The report focuses on Chinese AI companies' access to compute, as well as their distillation of American models, and how these factors translate into the current capability gap between the two countries.

Confirmed

• Measured by the Epoch Capabilities Index, China's top models currently lag US frontier models by about 6.5 months, meaning the US lead over China in AI is only around 6 months.
• If the federal government leveraged its existing policy authorities, it could widen that lead to more than a year; the report identifies two policies that could double the US advantage.
• The report includes an interactive analytical model that visualizes how the US lead would change under different policy interventions.

Why it matters

• The report advances the US-China AI competition debate from vague qualitative discussion to quantified assessment, directly simulating the effects of policy tools and providing a data-driven basis for US AI policy debates on China.
• It also warns that if existing policy tools are fully utilized, the US could more than double its lead—which could shape future discussions around export controls and compute restrictions.

2026-10-20 ~ 2026-10-20 · 7 related posts

• [source] IFP report: US AI lead over China is only ~6.5 months, and existing policy could double it — fiiiiiist · 2026-10-20
• [source] IFP Releases 23k-Word Report on How Compute Access and Distillation Shape the US-China AI Gap — jeremiecharris · 2026-10-20
• [source] IFP lists five export-control measures to double America's AI lead by end of 2027 — fiiiiiist · 2026-10-20

4 near-duplicate retellings: fiiiiiist · fiiiiiist · fiiiiiist · fiiiiiist