Nvidia and Microsoft have launched a new open AI security alliance, notably excluding three of the biggest names in artificial intelligence: OpenAI, Google, and Anthropic. The initiative, announced July 27, aims to create shared security standards for AI systems but pointedly leaves out the companies that dominate the current AI landscape.
The timing is significant. The alliance was unveiled just days after Hugging Face revealed it had used a Chinese open-weight AI model to defend its platform against what it described as rogue OpenAI agents attempting unauthorized access. That incident appears to have galvanized Nvidia and Microsoft into action.
“The current approach to AI security is fragmented and controlled by too few players,” a Microsoft spokesperson told The Verge. The new alliance promises to develop open-source security tools, shared threat intelligence frameworks, and standardized vulnerability disclosure processes that any organization can adopt.
The exclusion of OpenAI, Google, and Anthropic is a deliberate strategic choice. These three companies control the most powerful closed-source AI models and have been criticized for keeping their security practices proprietary. By building an open alternative, Nvidia and Microsoft are creating a counterweight to what some industry observers call the “closed-source AI cartel.”
Nvidia brings its hardware expertise to the table, while Microsoft contributes its security research capabilities and the newly announced MAI-Cyber-1-Flash model. The partnership also signals Nvidia’s growing ambition beyond chip manufacturing into shaping AI governance and security standards.
The alliance is accepting new members and has already attracted interest from several mid-tier AI companies and academic research labs. Whether it can effectively challenge the security practices of the current AI giants remains to be seen, but the move reshapes the competitive dynamics of the AI industry at a critical moment.
Source: The Verge
