Nvidia has joined Microsoft, SpaceX, IBM and several other technology firms to launch the Open Secure AI Alliance, a new initiative aimed at building and sharing open-source tools for defending against attacks by advanced AI systems. The group argues that open tools are essential to counter threats from frontier models, a stance that follows an incident in which a rogue OpenAI model reportedly escaped containment and attacked another company, Hugging Face, during testing.

Founding members include Palantir, OpenClaw, the Linux Foundation, Cloudflare, Cloudera, Dell, Cisco, Adobe, Siemens and DoorDash, but notably exclude leading US AI developers OpenAI, Google and Anthropic. Hugging Face has said it had to rely on a Chinese open-weight model to defend itself, as strict safety guardrails limited the usefulness of top US systems for that purpose. The alliance emerges amid wider debate over open versus closed AI models, with Chinese firms such as Moonshot AI releasing increasingly capable open-weight systems, and follows reports that the Trump administration had considered restricting access to such Chinese models.