The Pentagon has officially pulled the plug on Anthropic. Following months of speculation and internal friction, the Department of Defence confirmed to the BBC that it has finally ceased all use of Anthropic’s AI products. This marks the end of a messy, high-stakes standoff between the U.S. government and one of the biggest names in artificial intelligence.
The Fallout
Back in February, Secretary of Defence Pete Hegseth labeled Anthropic a "national security supply-chain risk"—a status usually reserved for foreign adversaries. The beef started when the Pentagon allegedly demanded Anthropic strip away its safety guardrails to give the military total, unrestricted access to its models. Anthropic pushed back, citing the real-world dangers of autonomous weapons and mass surveillance. The company actually sued the Trump administration over the designation, calling it unlawful.
Deep Integration Issues
Even though Hegseth set a hard deadline for the phase-out back in August, people in the know claim the Pentagon kept using Anthropic’s Claude models for intelligence gathering and military ops—including missions against Iran—as recently as last week.
Why the delay? Claude was baked deep into Maven, the Pentagon’s primary data platform operated by Palantir. As Lauren Kahn, a senior research analyst at Georgetown, noted, these massive AI systems aren't just plug-and-play. “Once they become integrated, it can be painful to remove them,” she explained.
The Plot Thickens
While the Pentagon says it's officially over, the vibe between the White House and Anthropic is lowkey shifting. President Donald Trump and Anthropic boss Dario Amodei met twice at the White House last week, with Trump publicly praising Amodei as “fantastic.” Meanwhile, the Pentagon is already pivoting, signing contracts with tech giants like Google, xAI, and OpenAI to fill the void.
Why it matters
This split is a major L for Anthropic’s government business, but the fact that the Pentagon relied on their tech for top-level intelligence shows exactly how vital AI has become for national defense. As the military swaps vendors, it proves that in the current AI arms race, the tech stack is the new battlefield—and it's giving major “complicated relationship” energy.
