Nvidia and 36 other technology companies announced the Open Secure AI Alliance on Monday, an effort to build open-source security tools for artificial intelligence systems. The group argues that defenders need the ability to run and inspect AI models on their own infrastructure—something closed, proprietary models often prevent.

The alliance includes Microsoft, IBM, Red Hat, Cloudflare, CrowdStrike, Palantir, Databricks, Hugging Face, and the Linux Foundation. Notably absent are OpenAI, Anthropic, and Google, developers of some of the most widely used closed models. The group's founding premise: when an AI system is under attack, delays caused by restricted access to model internals or dependence on external APIs can be the difference between containing a breach and suffering a full compromise.

Members are contributing specific open-source code. Nvidia released NOOA, a framework for testing and auditing AI agent behavior. Microsoft contributed MDASH, a multi-agent system designed to automatically discover exploitable software bugs. The Linux Foundation will oversee governance under its existing security initiatives. The alliance arrives as cybersecurity teams contend with increasingly sophisticated attacks on crypto networks and other high-value targets, where a single successful exploit can drain millions in irreversible transactions.

The push is timely. Recent incidents have highlighted the difficulty of forensic analysis when security teams lack the ability to run AI models locally. By keeping tools open, the alliance hopes to give defenders the same speed and adaptability as attackers—without relying on third-party APIs or closed model providers. Whether that approach gains traction among enterprises remains to be seen, but the lineup suggests broad industry support.

Source: CoinDesk