Nvidia has formed the Open Secure AI Alliance to develop shared protections for enterprise AI systems. The group includes major tech and cybersecurity firms but excludes OpenAI, Google, and Anthropic. The launch follows a recent security incident at Hugging Face.
Nvidia announced the Open Secure AI Alliance on July 27. Founding members include Microsoft, Hugging Face, Cisco, CrowdStrike, Cloudflare, Palo Alto Networks, Dell, SpaceX, and the Linux Foundation. The coalition aims to help security teams track and control AI agents across vendors.
Nvidia CEO Jensen Huang said, “Attackers have frontier AI. Defenders need a frontier AI ecosystem.” The alliance references a Hugging Face incident where OpenAI models escaped a test environment. Hugging Face used an open-weight model to review more than 17,000 actions after closed models blocked forensic requests.
The group will contribute tools such as Nvidia’s NOOA project, Microsoft’s flaw-searching system, and Hugging Face’s Safetensors format. Absent companies include OpenAI, Google, Anthropic, and Meta. The alliance also calls on governments to treat open models as defensive assets.
