Source: BBC

The Pentagon has finalized agreements with eight technology companies — Google, OpenAI, Amazon, Microsoft, SpaceX, Oracle, Nvidia, and the startup Reflection — to use artificial intelligence technologies for any lawful operational use by the US military. The department stated that access to a variety of AI capabilities across multiple suppliers aims to equip warfighters with effective tools and reduce reliance on any single vendor. Since its launch last year, the military’s AI platform has been used by over one million defence personnel to shorten task durations from months to days.

Anthropic is not part of these new contracts and was labelled a “supply chain risk” by U.S. Defence Secretary Pete Hegseth after concerns raised publicly by Anthropic’s CEO about AI uses in mass domestic surveillance and autonomous weapons. Anthropic is pursuing a legal challenge expected to be heard in court in September. Earlier in the year, OpenAI signed a contract with the Pentagon, and Google’s Gemini AI is now in use at the classified level for the first time. Nvidia and Reflection will supply open-source AI models, with Nvidia not providing hardware. Hundreds of Google and DeepMind employees have expressed internal opposition to expanding government collaborations involving AI.