OpenAI Chief Executive Officer Sam Altman warned in an 18 August interview with CNBC that Washington may be underestimating the speed and breadth of China’s advance in artificial intelligence. “I’m worried about China,” he said, adding that Beijing could outpace the United States in building the large-scale inference capacity needed to deploy frontier models. Altman argued that the current patchwork of U.S. export controls on semiconductors is unlikely to slow China’s progress. The Biden administration tightened curbs in 2024, and President Donald Trump in April 2025 temporarily halted all shipments of advanced GPUs before allowing sales of so-called ‘China-safe’ parts in July under a deal that forces Nvidia and AMD to remit 15 % of their China chip revenue to the U.S. government. Reports of a thriving black market—estimated at roughly US$1 billion in smuggled Nvidia processors during the three months after the April ban—underscore the limits of the policy, Altman said. Competition from Chinese labs is already shaping corporate strategy, Altman added, citing open-source systems such as DeepSeek as a key reason OpenAI released its first open-weight models, gpt-oss-120b and gpt-oss-20b, earlier this month. He described the U.S.–China contest as multidimensional—spanning research, product development and manufacturing—and called for a broader response than export controls alone: “You can export-control one thing, but maybe not the right thing… people build fabs or find other workarounds.”