Trump blacklists Anthropic. OpenAI steps in with Pentagon deal
Pentagon bans AI firm for refusing to remove guardrails on surveillance and weapons
Anthropic signed a contract worth up to $200 million with the Pentagon in July 2025 to 'advance responsible AI in defense operations.' The company worked with data analytics firm Palantir to deploy its Claude AI model on classified defense and intelligence networks. Claude was the first AI model cleared to operate on classified military systems.
Key Concepts
Essential concepts and terms to understand this topic
Autonomous Weapons Systems
Military weapons that can select and engage targets without direct human intervention, raising questions about accountability and compliance with international law.
Artificial Intelligence Ethics
The set of principles — fairness, accountability, transparency, and safety — that govern how AI systems should be designed and deployed.
Regulation
Rules issued by federal agencies that implement statutes and carry the force of law.
AI Regulation
Government rules governing how artificial intelligence systems are developed, tested, and deployed, including safety requirements and use restrictions.
Oversight and Accountability
The principle that government officials and institutions must answer for their actions.
Key People
9DA
Dario Amodei
CEO and Co-Founder, Anthropic
Amodei met with Hegseth in a high-stakes Pentagon meeting Tuesday before the deadline. He publicly refused the Pentagon's demands, saying Anthropic 'cannot in good conscience' allow Claude to be used for autonomous weapons or mass surveillance. He called the restrictions 'two narrow exceptions' that had never blocked a single government mission.
Pete Hegseth
U.S. Secretary of Defense
Hegseth set the 5:01 PM deadline and immediately announced the supply-chain risk designation when it passed. He said 'America's warfighters will never be held hostage by the ideological whims of Big Tech' and called the decision 'final.' He simultaneously had undersecretary Michael privately offering Anthropic a deal.
EM
Emil Michael
Under Secretary of Defense for Research and Engineering
Michael led Pentagon negotiations with Anthropic and other AI firms. He publicly called Amodei a 'liar' with a 'God complex' on X. Sources told Axios he was on the phone offering Anthropic a backdoor deal at the same moment Hegseth publicly announced the blacklist — a deal that would have permitted AI-enabled collection of American personal data.
SA
Sam Altman
CEO, OpenAI
Altman wrote in a Thursday staff memo that OpenAI shares Anthropic's red lines on surveillance and autonomous weapons. Yet hours after Anthropic was blacklisted, Altman announced OpenAI's own Pentagon deal, saying the DoW agreed in the contract not to use OpenAI models for domestic mass surveillance or autonomous weapons. Critics questioned whether the terms were meaningfully different from what Anthropic was denied.
Donald Trump
President of the United States
Trump ordered every federal agency to 'immediately cease' using Anthropic products via Truth Social, calling the company a 'radical left, woke company.' He threatened civil and criminal consequences and gave agencies a six-month phase-out period. His post framed the dispute as the government's right to decide how military AI is used.
Mark Warner
U.S. Senator, D-VA; Vice Chair, Senate Select Committee on Intelligence
Warner condemned Trump's action as potentially steering contracts to a preferred vendor and questioned whether national security decisions were driven by analysis or politics. He noted xAI, Elon Musk's AI company, had just received the second classified network clearance, suggesting political favoritism.
Sean Parnell
Pentagon Chief Spokesperson
Parnell posted on X that the Pentagon's request was a 'simple, common-sense' requirement to prevent Anthropic from 'jeopardizing critical military operations.' He framed Anthropic's restrictions as potentially putting warfighters at risk, escalating the public pressure campaign before the deadline.
GA
Greg Allen
Senior Advisor, Center for Strategic and International Studies (CSIS)
Allen told CNN that Claude's usage restrictions 'have never been triggered' in conversations with Defense Department users, undercutting the Pentagon's argument that Anthropic's guardrails were causing operational problems. He called the standoff 'highly unusual' in the world of Pentagon contracting.
AJ
Andy Jassy
CEO, Amazon
Amazon announced a $50 billion investment in OpenAI on February 27, the same day Anthropic was blacklisted. Amazon is a major investor in Anthropic and also provides AWS cloud services to both companies, creating a complex conflict of interest as the Pentagon pivoted away from Anthropic toward OpenAI.