Meta, Microsoft, Nvidia, IBM, and others back open-weight AI

Details

A coalition of more than 20 companies and organizations published an open letter today urging US policymakers to preserve access to open-weight AI models — systems where trained parameters are publicly downloadable, modifiable, and deployable on private infrastructure. Signatories span direct commercial rivals: Meta, Microsoft, Nvidia, IBM, Dell Technologies, CrowdStrike, Palantir, ServiceNow, Hugging Face, Perplexity, Mistral, Andreessen Horowitz, Y Combinator, the Linux Foundation, and Mozilla, among others. The letter makes three core arguments: open weights lower the barrier to entry for startups and institutions that cannot afford frontier model training costs or per-token API fees; they sustain competition across the AI stack from chips to applications, keeping prices down; and they give enterprises genuine data sovereignty by allowing on-premise deployment without dependence on a single vendor roadmap or pricing. The letter addresses security head-on, conceding that once weights are released they cannot be recalled and safety guardrails can be stripped. The counterargument draws on cybersecurity: defenders need open access to the technology they protect against, and restriction alone does not eliminate risk. The timing appears deliberate. The Trump administration has reportedly been weighing action against Chinese open-weight models and distillation practices, and the letter explicitly defends distillation as a legitimate, historically grounded form of technical learning.