The Pentagon has signed AI contracts with Nvidia, Microsoft and Amazon - but not Anthropic

The U.S. Defense Department has entered into agreements with seven artificial intelligence companies to deploy their advanced technologies on its classified networks. The Pentagon is seeking to expand the range of AI vendors working with the US military, Reuters explains. At the same time, it continues to ignore one of the leaders in the AI field - the startup Anthropic.

The list of new suppliers includes SpaceX-owned xAI, startup OpenAI, Google, Nvidia, AI startup Reflection, Microsoft and Amazon Web Services. Some of these companies are already working with the Pentagon, Reuters notes. They will all be integrated into its classified networks, giving the military greater access to their products to handle sensitive topics, the Pentagon said in a statement.

"These agreements accelerate the transformation to make the U.S. Army a warfighting force for which AI comes first, and will strengthen our military's ability to maintain decision-making superiority in all areas of warfare," the Pentagon said.

The statement did not mention AI startup Anthropic. The company had previously restricted the Pentagon from using its software for ethical reasons. The Pentagon then recognized the AI startup as a supply chain risk, banning its use by its contractors.

By expanding the list of AI services available to the military, the Pentagon, according to the statement, intends to avoid being "tied to a single vendor." This is likely a hint at the agency's over-reliance on Anthropic or other dominant service providers, Reuters notes.

Anthropic's Claude model has previously been the only one the military has used in covert environments, including in the war against Iran and in the capture of Venezuelan leader Nicolas Maduro, says the Financial Times. However, Anthropic CEO Dario Amodei said he would not sign an agreement with the Defense Department unless the government could guarantee that Anthropic's tools would not be used for mass domestic surveillance and autonomous weapons. In the wake of the Pentagon's conflict with Anthropic, new AI players say the military has accelerated the process of connecting them to classified data - it now takes less than three months, Reuters writes. Previously, the process took 18 months or more, the agency specifies.

This article was AI-translated and verified by a human editor