Defense contractors and federally funded labs are unclear on whether they can keep using Anthropic’s Claude AI without risking their government work.
Defense Secretary Pete Hegseth on Feb. 27 designated the company a national security supply-chain risk, a label typically reserved for foreign adversaries, and ordered that “effective immediately, no contractor, supplier, or partner that does business with the United States military may conduct any commercial activity with Anthropic.”
The move followed Anthropic’s refusal to let the Pentagon use Claude for fully autonomous weapons or mass domestic surveillance of Americans. President Trump on Truth Social directed “every Federal Agency in the United States Government to immediately cease all use of Anthropic’s technology” and ordered a six-month phase-out for military users.
CEO Dario Amodei told CBS the label will restrict contractors from using the model on any projects touching their military contracts. Compliance teams are already mapping what “touches” actually means for lab automation, literature reviews and code pipelines.
The reason for the squabble? Anthropic CEO Dario Amodei refused to budge on two “red lines” for military use of Claude: First, no mass surveillance: AI capabilities could enable surveillance at a scale previously impossible, including government purchase of private data for analysis. Second, no autonomous weapons. Amodei said AI isn’t reliable enough, and accountability for lethal decisions is unclear. Amodei said: “We don’t want to sell something that we don’t think is reliable.”
The Pentagon wanted Claude access for “all lawful purposes.” Pentagon CTO Emil Michael argued existing federal law already restricts the disputed uses and said “At some level, you have to trust your military to do the right thing.” The Pentagon offered written acknowledgments of the restrictions, but Anthropic rejected them as containing legal loopholes.
The public backlash has cut the other direction. Claude hit No. 1 on Apple’s U.S. App Store Top Downloaded chart during the standoff.
For R&D organizations, the immediate question is less about whether Claude is “banned” outright than about how broadly the restriction is interpreted and enforced across prime contractors, subcontractors and federally funded labs. Even if a company only uses Claude for mundane work like summarizing papers, drafting test plans or helping engineers debug code, the models themselves are widely viewed as top-tier for software work, including top placements on public coding leaderboards and strong results on SWE-bench Verified, the software engineering benchmark.
Compliance officers now have to decide whether any of that activity could be viewed as “commercial activity” with Anthropic or as work that indirectly supports a covered program. That is pushing teams toward interim workarounds like freezing new Claude deployments, tightening access controls, separating accounts and data environments by contract, and accelerating evaluations of alternative models that can be hosted or procured under existing government-approved pathways.
In Reddit subbreddits, the framing is mostly moral and consumer-brand driven, not “what do DFARS clauses mean for my pipeline.” One post, for instance, explicitly pitches the situation as people “quitting ChatGPT in favor of Claude” because of surveillance and weapons concerns. But opinions are mixed. One commenter says anyone who thinks there’s a “good guy among AI companies” needs “a huge reality check.” Another calls it “performative nonsense” and mocks “allegiances” to huge companies “like it’s a team sport” while other comments point to the popularity of Anthropic models. Another commenter flatly says “Claude AI is top dog” for their workflow. The company’s coding tools are among the most popular in the industry, with Claude models routinely ranking at or near the top of leaderboards such as Arena.ai based on Elo scores.
Amodei in the CBS interview said Anthropic is “still interested in working with them as long as it is in line with our red lines,” and characterized disagreement with the government as fundamentally American, arguing his company has acted patriotically and in national security interests.
