Pentagon Signs Deals with Six AI Firms for Classified Networks
The Pentagon revealed on Friday that it secured agreements with leading technology firms to boost the military's artificial intelligence functions. These pacts authorize the companies to join classified networks, expanding access for defense operations.
The six companies committed to letting the Defense Department apply their technology in any lawful manner. This clause has faced pushback from Anthropic, which stood as the sole AI model accessible on classified systems before these developments.
Earlier confirmations included arrangements with Elon Musk's xAI, OpenAI, and Google. New additions cover Amazon Web Services, Microsoft, Nvidia, and Reflection AI, a startup focused on advanced AI applications.
Agreements Advance Military AI Goals
A Pentagon statement described these deals as speeding up the shift to position the United States military as an AI-first fighting force. Officials did not detail specific uses for the tools but noted benefits for service members in reaching quicker, superior decisions.
The department emphasized that a varied range of AI options from the strong American technology base equips warfighters with essential resources. This setup allows confident actions to protect the nation from threats.
Defense leaders expect these pacts to encourage Anthropic to abandon its hesitations regarding the broad any lawful use policy.
Ongoing Tensions with Anthropic
Stay ahead of the AI curve
The most important updates, news, and content — delivered weekly.
No spam. Unsubscribe anytime.
President Trump directed the government to sever connections with Anthropic. Yet the company's technology persists on classified networks, and intelligence analysts continue relying on its models. Efforts to switch rapidly to OpenAI's ChatGPT and Google's Gemini have encountered hurdles and technical issues.
Anthropic and the Pentagon face federal court battles over the Defense Department's designation of the firm as a supply chain risk. This marks an unusual application of government authority to question corporate product development practices.
White House staff, both admiring and concerned about Anthropic's latest model, Mythos, seek a resolution to the conflict. Such a deal could enable other government sectors to collaborate with the company.
Bloomberg first reported the agreements with these companies. A Pentagon official noted that the new terms avoid vendor lock, ensuring the military avoids dependence on a single provider. The department prefers a uniform standard and resists providing firms with assurances on model usage.
Anthropic and the Pentagon disagree on deploying the company's Claude model for piloting autonomous drones or domestic surveillance tasks. The Pentagon asserts no plans for such uses, yet the parties lack accord on contract wording or its necessity.
Company and Reporter Background
These developments occur against the backdrop of Anthropic, founded in 2021 by former OpenAI executives with a focus on safe AI systems. OpenAI, creator of ChatGPT, leads in generative AI. Google, part of Alphabet Inc., developed Gemini as a competitor model. xAI, launched by Elon Musk, aims to understand the universe through AI. Amazon Web Services provides cloud infrastructure key to AI training. Microsoft integrates AI across its products, including partnerships with OpenAI. Nvidia supplies critical chips for AI computations. Reflection AI, a newer entrant, specializes in agentic AI for real-world tasks.
Reporters Julian E. Barnes, who covers U.S. intelligence and security for over two decades, and Sheera Frenkel, based in the San Francisco Bay Area focusing on technology's societal impacts including social media platforms, detailed the story.