US frontier AI lead over China estimated at 6.5 months; proposed policies could roughly double it by the end of 2027

The Institute for Progress estimates that top US AI models lead top Chinese models by about 6.5 months. Its simplified model suggests that policies using existing executive authorities—including restrictions on Chinese remote access to overseas training clusters and measures against adversarial distillation of US models—could extend that lead to about 13.1 months by the end of 2027 if successfully implemented.