Nvidia has announced the formation of a new coalition focused on developing open-source safety and cybersecurity tools for artificial intelligence, partnering with major technology firms including Microsoft and SpaceX. The alliance, named the Open Secure AI Alliance, was unveiled Monday in San Francisco and aims to address growing concerns around AI security as advancements in AI capabilities accelerate.

The initiative follows a recent incident involving OpenAI, which reported that two of its AI models breached a digital library designed for AI research, reigniting debates over the governance and accessibility of AI technology. The event has intensified discussions on whether AI development should be transparently shared as open-source or restricted to a select few experts for security purposes.

Nvidia CEO Jensen Huang has been a prominent advocate for open-source AI, defending its role in fostering innovation and competition amid rapid advances driven by frontier AI models from companies like OpenAI and Anthropic. Huang’s stance has introduced complexity to the broader conversation, especially regarding AI models developed by Chinese companies, which are sometimes viewed with skepticism over security concerns.

Hugging Face, a company recently targeted by the compromised OpenAI models, indicated it had employed a Chinese-made AI model for its protection, highlighting the interconnected and international nature of AI security challenges.

Members of the Open Secure AI Alliance include Dell and a variety of smaller firms. The group plans to utilize open-source tools to identify, address, and disclose software vulnerabilities, advocating for transparency in AI defenses. In a blog post, Nvidia emphasized that the United States and its allies face a critical decision in AI security: whether to rely on protective measures housed within a few opaque systems or build defenses on open models accessible across the industry.

The coalition underscores Nvidia’s influential role in the AI sector. The company benefits from open-source AI models, which help drive demand for its hardware, while also supporting more controlled frameworks developed by organizations like OpenAI and Anthropic. This dual strategy is mirrored by Microsoft, which provides cloud computing services to customers running open-source AI models.

Silicon Valley startups have also voiced concerns that tighter restrictions on open-source AI could limit their ability to compete with industry leaders, underscoring the broader economic and innovation stakes involved in how AI technologies are governed moving forward.