The U.S. Department of Defense has struck major deals with seven leading technology companies to deploy their artificial intelligence tools across classified networks, a move announced on Friday that pointedly excludes Anthropic amid ongoing disputes.
This development marks a significant shift for the Pentagon, which aims to build an “AI-first fighting force and will strengthen our warfighters’ ability to maintain decision superiority across all domains of warfare.”
Officials stressed that the AI systems will support only “lawful operational use,” as outlined in the department’s release. The platform GenAI.mil has already drawn 1.3 million users from DoD ranks, demonstrating early momentum in military AI adoption.
The partners comprise Elon Musk’s SpaceX, OpenAI (makers of ChatGPT), Google, Microsoft, Nvidia, Amazon Web Services, and Reflection. These build on existing ties with firms like Palantir and OpenAI.
By contrast, Anthropic finds itself frozen out after clashing with the Trump administration over safety requirements for its Claude model in combat scenarios, including autonomous weapons and surveillance. Until recently, Claude was the sole AI authorised for such secure environments.
The rift deepened when President Trump moved to sever ties because Anthropic refused terms allowing use for “all lawful purposes.” The Pentagon branded it a “supply chain risk,” a severe label typically reserved for foreign adversaries, prompting a lawsuit from the company. A California federal judge blocked the restriction last month.
Recent overtures suggest possible reconciliation. Anthropic CEO Dario Amodei met White House Chief of Staff Susie Wiles last month, coinciding with the launch of the firm’s Mythos tool for cybersecurity threat detection, which also highlights potential vulnerabilities.
Fuelled by £12.5 billion from last year’s One Big Beautiful Bill Act for AI and cyber operations, the deals position the administration to pressure Anthropic, valued at $60 billion, while rivals chase lucrative contracts.
Critics from groups like the Future of Life Institute caution that bypassing safety guardrails risks hastening unregulated lethal AI, echoing NATO concerns. As the AI arms race intensifies, the Pentagon’s strategy promises operational speed but spotlights enduring tensions between technological prowess and ethical boundaries.