Nvidia Forms 60-Member Open Secure AI Alliance to Fight Back Against AI-Powered Attacks
Microsoft, SpaceX, Cisco, CrowdStrike, and dozens of others join Nvidia's new coalition for open-source cyber-defense tools, launched days after defenders leaned on an open-weight Chinese model when closed AI tools blocked their forensic work.
Nvidia launched the Open Secure AI Alliance on July 27, bringing together more than 60 founding organizations — including Microsoft, SpaceX, Dell, Cisco, CrowdStrike, IBM, Palo Alto Networks, Red Hat, Hugging Face, and the Linux Foundation — around a shared goal: build and share open-source tools for AI-driven cyber defense. The alliance builds on the Linux Foundation’s existing open-source security work rather than starting from scratch.
The stated rationale is that defenders need access to open, trustable frontier tools instead of relying solely on closed commercial models controlled by a handful of vendors. That’s a notable position for Nvidia to stake out publicly, given how much of the current AI boom runs through proprietary models from labs it also invests in.
The timing isn’t coincidental. The alliance’s launch came days after a widely reported incident in which security researchers investigating a compromise turned to a self-hosted open-weight Chinese model, GLM 5.2, to analyze more than 17,000 malicious actions after closed commercial AI tools blocked their forensic work — reportedly due to policy restrictions on the systems being examined. That episode became a concrete argument for exactly what the Open Secure AI Alliance is proposing: defenders can’t be dependent on a vendor’s content policy deciding, in the middle of an active incident, what they’re allowed to analyze.
It’s also a defensive move in a market where AI is increasingly a two-way weapon. Attackers have been using generative models to write more convincing phishing content, automate reconnaissance, and generate polymorphic malware faster than signature-based defenses can keep up. Security teams need equivalent AI-assisted tooling to keep pace, and a coalition this size — spanning cloud infrastructure, endpoint security, hardware, and open-source foundations — signals an industry-wide bet that the tooling gap needs to close through shared, inspectable code rather than another round of closed vendor products.
Whether the alliance produces usable tools quickly enough to matter is the real test. Coalitions this size move slowly by nature, and the security teams dealing with AI-accelerated attacks today need defenses now, not a multi-year standards process. Nvidia’s involvement at least guarantees hardware-level backing and hosted compute for whatever tooling emerges — the harder problem is coordinating 60-plus organizations toward shipped, maintained software rather than a press release with a long member list.