Silicon Valley’s rift over open source pushes back contemplated White House bans on Chinese AI

Details

According to a New York Times report, White House Chief of Staff Susie Wiles, Treasury Secretary Scott Bessent, and Commerce Secretary Howard Lutnick spent weeks deliberating aggressive intervention in the open-source AI market. Options on the table included: sanctions against Chinese open-source AI model makers, trade blacklisting, bans on US cloud companies serving those firms, and Commerce Department rules restricting which models federal agencies may use. The divide was sharp and commercially motivated. OpenAI and Anthropic — both closed-model companies with planned IPOs — pushed for restrictions while citing national security. NVIDIA, Microsoft, Google, and Meta, which all distribute or benefit from Chinese open-source models, coordinated a counter-argument through private calls and video conferences. Jensen Huang wrote his first-ever X post on July 24 to publicly defend open models, and helped grow the Open Secure AI Alliance from a few members to over 230. The precipitating event was Kimi K3 from Moonshot AI, which the White House OSTP head Michael Kratsios publicly accused of distilling Anthropic's Fable model. Treasury Secretary Bessent floated sanctions the same day. The administration has since shifted to 'making US models more competitive' rather than outright bans — for now. A policy decision is expected before Xi Jinping's September visit. The outcome will materially reshape the competitive landscape for every company building on or competing with Chinese open-source AI.