The AI race between the United States and China is described as infinite, unlike the finite 20th-century space race. The United States appears ahead, but the article suggests a lack of a classified national strategy could become a global leadership crisis. Export controls for GPUs and sanctions are deemed insufficient, as China has shown near-latest GPUs can be potent. A bigger problem identified is that optimization approaches at Anthropic, OpenAI, Meta, and DeepMind are similar, all expanding through existing architectures rather than detouring into basic research. The article recommends that U.S. National Laboratories, including Argonne, Oak Ridge, Los Alamos, Lawrence Berkeley, Lawrence Livermore, Sandia, and Savannah River, pursue fundamental research to achieve artificial general intelligence with superintelligence capabilities. It names Michael Kratsios, Director of the Office of Science and Technology Policy; Chris Wright, Secretary for the US Department of Energy; and Emil Michael, Under Secretary for Research and Engineering, as officials who should focus on more than regulation.