SpaceX joins Nvidia's AI security bloc — OpenAI sits it out
50+ firms launched the Open Secure AI Alliance under the Linux Foundation. OpenAI, Google and Anthropic aren't in it.
SpaceXAI signed on as a founding member of the Open Secure AI Alliance, the Linux Foundation–governed group NVIDIA announced on 27 July 2026 with 50+ inaugural partners: Microsoft, IBM, Cisco, Cloudflare, CrowdStrike, Hugging Face, Palo Alto Networks, Red Hat, Databricks, Mistral, GitHub and more.
The dowry matters more than the logo. SpaceXAI open-sourced Grok Build, its terminal coding agent, and says it will release weights for the Grok model line. NVIDIA is contributing open models, weights, data and NOOA (Object-Oriented Agent), an Apache 2.0 framework now on GitHub for making agent behaviour testable, traceable and auditable — with a blunt caveat in the repo that its controls are "not a containment boundary," so agent-generated code still needs container or VM isolation.
The alliance builds on OpenSSF and the Linux Foundation's Akrites project, and lands just after a July intrusion in which an autonomous agent reached parts of Hugging Face's production infrastructure via a malicious dataset carrying a remote-code loader. Conspicuously absent from the roster: OpenAI, Google, Anthropic and Meta — the labs holding the most closed frontier weights.
Why it matters: agent security is turning into a standards fight, and the open-weight camp just organised first.
Sources
Primary: the company, paper or repository
Independent coverage
Written by an AI pipeline from the sources above. Methodology · Report an error
Feed, daily deep-dive and bytes — readable offline, with push alerts for the topics you follow.