Secretary Pete Hegseth is trying to destroy a leading AI company for no good reason.
Pete Hegseth has made good on his threat to destroy leading AI firm Anthropic, using a little-used authority to designate it as a “supply-chain risk.” It’s a boneheaded move that won’t make supply chains any safer, but will harm American interests and deter investment in new technologies.
The Wall Street Journal reports:
The Defense Department on Thursday formally notified Anthropic’s leadership that the company and its AI tools present security threats that require an action normally reserved for businesses from foreign adversaries, a senior Pentagon official said.
The official said the military wouldn’t allow a vendor to insert itself into the chain of command and put members of the armed services at risk by restricting the lawful use of a critical capability.
The spat between Hegseth and Anthropic began when the company refused to accede to the Pentagon’s unconditional use of its AI model Claude. (Anthropic would not let Claude be used for mass domestic surveillance or autonomous killing without human input, while the military demanded full access to “all lawful purposes.”) Hegseth could have stopped at canceling Anthropic’s $200 million contract with the Defense Department and switching to another AI provider. It did exactly that last week, signing a new deal with OpenAI for its ChatGPT model that — get this — included the same ethical-safeguard provisions that caused the Pentagon to drop Anthropic.
But the administration couldn’t leave it at that. President Trump ordered on social media that all federal agencies stop using Claude, which they had been integrating into their operations for months. (Easier said than done, as the military relied upon Claude to assist its Iran strikes hours later.) The supply-chain risk designation is far more sweeping: Hegseth has said that it would prohibit any defense contractor from doing business with Anthropic.
Turns out, a lot of contractors already do business with Anthropic. Weapons suppliers like Lockheed Martin use Claude for everyday tasks. Three tech giants that provide the military with cloud services — Amazon, Google, and Microsoft — are major investors in Anthropic. Now, defense-tech companies are cutting their ties with Anthropic in droves to stay in the Pentagon’s good graces.
Everyone, including Hegseth, knows that Claude is not a genuine risk to military supply chains. To label Anthropic as such — putting it on the same level as Chinese military companies — is a joke. Speaking of China, the PRC must be thrilled that America is kneecapping one of its most advanced AI companies.
The damage is not limited to Anthropic. Tech companies may now think twice before working with the U.S. military — or developing new technologies with defense applications in the first place. The upside is that you may get a lucrative contract. The downside is that the defense secretary might destroy your entire business on a whim if he loses his temper.
Anthropic has promised to challenge the supply-chain risk designation in court. It should win that case on both constitutional and statutory grounds. When it does, the company will be doing America a favor.