Nvidia has pulled together a 37-member AI security alliance, bringing together chipmakers, cloud providers, and cybersecurity firms — but notably leaving out OpenAI, Anthropic, and Google. The group, announced Tuesday, is built around a simple premise: defenders need AI they can run themselves, not just rely on tools from a handful of dominant AI labs.
Who's in and who's out
The alliance includes companies like Arm, Cisco, Dell, Hewlett Packard Enterprise, Lenovo, and Microsoft. Cloud providers Amazon Web Services, Google Cloud, and Microsoft Azure are all members. But the big AI model makers — OpenAI, Anthropic, and Google itself — are not part of the coalition. That absence is striking given that those three companies are at the center of the current AI boom and have their own security teams and products.
Nvidia's rationale, according to the company, is that the security community needs AI tools it can deploy and control independently. The alliance aims to give defenders — from large enterprises to smaller organizations — access to AI-powered security solutions that aren't tied to a single vendor's model or platform.
Why the absence matters
OpenAI, Anthropic, and Google are developing their own security AI, but they're also the companies whose models are most widely used. Their absence from the alliance means the group won't have direct input from the creators of GPT-4, Claude, or Gemini. That could limit the alliance's ability to address vulnerabilities in those specific models — or to shape how those companies approach security.
Still, Nvidia's move reflects a broader push in the industry: make AI security tools more accessible and less dependent on a few big players. The alliance's members cover a wide range of the tech stack, from chips to cloud to endpoint security. That breadth could help it develop standards and tools that work across different environments.
What the alliance will do
Details are sparse. The group says it will work on shared frameworks, reference architectures, and open-source tools for AI security. The focus is on helping organizations run their own AI security models — for tasks like threat detection, anomaly spotting, and automated response — rather than relying on external APIs or proprietary systems.
Nvidia is providing its AI Enterprise software and GPU infrastructure to support the effort. Other members are expected to contribute their own technologies and expertise. The alliance is open to new members, though the criteria for joining haven't been spelled out.
The timing is no accident. AI adoption is accelerating, and so are the risks — from data poisoning to model theft to adversarial attacks. Companies are scrambling to secure their AI deployments, but many lack the in-house expertise or the right tools. The alliance is an attempt to fill that gap, at least for those who can afford Nvidia's hardware.
Whether the absence of the biggest AI labs will hurt the alliance's credibility remains an open question. For now, Nvidia is betting that a broad coalition of infrastructure and security players can move faster — and more independently — than any single company.
