Trump orders federal agencies to stop using Anthropic AI; Hegseth labels it a "supply-chain risk"

#What happened

On Friday, October 16, the Pentagon and Anthropic failed to agree on terms for a $200 million contract by a 5:01 p.m. deadline. The military wanted the ability to use Anthropic's AI for "all lawful purposes"; Anthropic wanted contractual assurances its AI would not be used for domestic surveillance of Americans or for fully autonomous lethal weapons.

Defense Secretary Pete Hegseth then designated Anthropic a "supply-chain risk to national security," a label that cuts the AI company off from business with the U.S. government. President Trump ordered federal agencies to immediately stop using Anthropic's technology and called the startup a "radical Left AI company."

Hours later, OpenAI announced it had reached an agreement with the Pentagon to deploy its AI technologies on classified systems. Altman negotiated differently from Anthropic: OpenAI agreed to "all lawful purposes" use, but secured the right to build technical safeguards that prevent domestic surveillance and autonomous-weapons use into its own systems. The Pentagon agreed to embed some OpenAI employees alongside government personnel on classified projects "to help with our models and to ensure their safety."

Two other developments the same week: the Pentagon reached an agreement earlier in the week to use Elon Musk's xAI technology for classified operations, and Amazon signed a $50 billion investment into OpenAI as part of $110 billion in funding OpenAI raised. Meanwhile, dozens of OpenAI employees signed an open letter urging AI companies to hold a common line against domestic surveillance and autonomous-weapons use, and Altman said on CNBC on Friday that he mostly trusted Anthropic and that "they really do care about safety."

#Why it matters

The U.S. government's AI vendor landscape just split in two: Anthropic is now blacklisted from federal business over its red lines, while OpenAI took the classified-systems contract by embedding its own technical guardrails instead of contractual refusals. The Amazon and xAI deals show the Pentagon is stitching together multiple AI vendors for classified work at the same time. For enterprise AI buyers, the message is that government AI contracts now turn on who controls the safety stack, and for Anthropic the "supply-chain risk" label is an existential federal-sales threat.

Verified October 17, 2026.