NVIDIA Starts Open Secure AI Alliance for AI Security Tools

NVIDIA announced in a company blog post the Open Secure AI Alliance, a group focused on building and sharing open tools for AI safety and cybersecurity. The company said founding members include Adobe, Cisco, Cloudflare, CrowdStrike, Databricks, Dell Technologies, HPE, Hugging Face, IBM, Microsoft, Palo Alto Networks, Red Hat, Salesforce, SAP, ServiceNow, Siemens, Snowflake and others.

The alliance will work on vulnerability remediation and disclosure using open technologies, building on the Linux Foundation's Akrites initiative and OpenSSF community work. It will focus on tools for AI agents, including identity, permissions, harnesses, guardrails, logs and evaluation.

NVIDIA is contributing open models, model weights, data and agent harness research. Its NOOA research framework is available on GitHub and is designed to help harnesses test, trace, audit and govern agent behavior.

Other contributions include HPE work on SPIFFE and SPIRE, Hugging Face Safetensors, IBM and Red Hat Lightwell, Microsoft's MDASH scanning harness and SpaceXAI's Grok Build coding agent. The alliance also called for policymakers to treat open models, harnesses and security tooling as defensive assets in AI and cybersecurity policy.