NVIDIA launches Open Secure AI Alliance with more than 60 partners

5:00 AM ET · July 30, 2026

NVIDIA announced the Open Secure AI Alliance this week, an industry group of more than 60 inaugural partners spanning cloud computing, cybersecurity, enterprise software and AI research that plans to build and distribute open source tools for defending AI agents and the systems they run on. Members include Microsoft, Adobe, Box, Cisco, Cloudflare, CrowdStrike, Databricks, Dell Technologies, GitHub, HPE, Hugging Face, IBM, the Linux Foundation, Mistral, Palantir, Palo Alto Networks, Red Hat and Salesforce, among others. The alliance’s premise is that open, inspectable security tools give defenders capabilities they can control and trust, in the same way open models have broadened access to AI capability more generally. Founding technical contributions include NVIDIA’s own NOOA research framework, Hugging Face’s Safetensors format for securely storing model weights, Microsoft’s MDASH tool for scanning across multiple models, and xAI’s decision to open source its Grok Build coding agent. The alliance is framed as building an open defense stack for agents, covering identity verification, secure model formats and safe coding workflows, at a moment when incidents like the recent Hugging Face breach involving an OpenAI model have sharpened industry focus on agentic AI security.