Nvidia has pulled together a 37-member coalition focused on artificial intelligence security, a move that conspicuously leaves out two of the biggest names in the field: OpenAI and Google. The alliance, announced this week, aims to address vulnerabilities in AI systems and could push the industry toward more open-weight models.
Who's in, who's out
The group includes a mix of chipmakers, cloud providers, startups, and research institutions. Nvidia itself is the anchor. Other members span sectors from cybersecurity to enterprise software. But the absence of OpenAI and Google — both heavyweights in AI development — is notable. Neither company is part of the alliance, though the facts don't specify why. The exclusion could signal a divide between proprietary and open approaches to AI safety.
Why open-weight models matter
Open-weight models allow developers to inspect and modify the underlying parameters of an AI system. That transparency can help identify security flaws, but it also raises concerns about misuse. The alliance's focus on open-weight models may shift industry norms, making such models more common. That could change how companies approach AI security — moving from closed, proprietary systems toward collaborative, community-driven defenses.
Potential impact on cybersecurity and regulation
The alliance's work could ripple beyond individual companies. Cybersecurity strategies might adapt to account for open-weight vulnerabilities, and regulators could look to the group's standards as a template. With 37 members, the coalition has enough weight to influence policy discussions. But without OpenAI and Google, any resulting norms may face pushback from the largest AI players. The question now is whether the alliance can set a de facto standard — or if it will remain a niche effort.
No timeline for the group's first deliverables has been announced. The industry will be watching to see if more members join — and whether the excluded giants eventually come to the table.
