NVIDIA rallies industry giants to establish Open Secure AI Alliance

NVIDIA has launched the Open Secure AI Alliance alongside dozens of founding partners to build open models and tools for AI security, a move that could expand market opportunities for defenders worldwide.

The coalition brings together industry names including Adobe, Cisco, Cloudflare, CrowdStrike, Databricks, Dell Technologies, HPE, Hugging Face, IBM, LangChain, the Linux Foundation, Microsoft, NetApp, OpenClaw, Palantir, Palo Alto Networks, Red Hat, Salesforce, ServiceNow, Siemens, Snowflake, SpaceXAI, Thinking Machines Lab and TrendAI.

The alliance aims to democratize AI defense capabilities by providing open, frontier tools that organizations can deploy with trust, according to NVIDIA. The approach follows the open-source model that proved critical to modern cloud computing, financial services, telecommunications and government services, with cybersecurity ranking among the top three beneficiaries of accessible technology and knowledge.

NVIDIA framed the initiative against the backdrop of the recent Hugging Face security incident, in which a closed AI model blocked forensic analysis because it could not distinguish attackers from defenders, while the open model GLM 5.2 contained the intrusion. The company argues that organizations without adaptable models face constrained response capabilities at moments when speed matters most for business continuity.

Several founding members are already contributing technology. NVIDIA Labs' Object-Oriented Agent (NOOA) is an open-source project designed to make AI safety accessible for agent harnesses, enabling easier testing, tracing and governance of agentic behavior; HPE is backing open standards such as SPIFFE/SPIRE to strengthen Zero Trust identity for AI; Hugging Face has donated Safetensors to the PyTorch Foundation; and IBM and Red Hat's Lightwell adds digitally signed patches to secure the open-source supply chain.

Microsoft's MDASH uses multiple AI agents to uncover software vulnerabilities, and SpaceXAI has open sourced its Grok Build coding agent with plans to release Grok model weights. According to NVIDIA, security requires examination across the entire agent stack, including identity, permissions, harnesses, guardrails, logs and evaluation.

NVIDIA called on companies and governments to invest in shared open infrastructure for AI defense, arguing that the approach mobilizes the full community of defenders rather than relying on secrecy alone for safety. OpenSSL President Tim Hudson said open source gives organizations complete control over their own deployment and none whatsoever over anyone else's.

What to watch: how the alliance's open tools perform against closed alternatives in real incidents, whether more players join the coalition, and how governments respond to the call for shared open infrastructure for AI security.