Anthropic wanted Pentagon business, and now the Pentagon is dangling a supply-chain designation over its head. The weird part is how shaky the legal machinery looks on both sides. Welcome to Anthropic Pentagon Watch. Today we’re digging into what Hegseth’s "supply chain risk" framing actually lets DoD do, and what it absolutely does not. And while the lawyers argue over designation thresholds, the other AI companies are still signing DoD deals. So the real question is who slides into the contract gap if Anthropic gets frozen out. We’ve also got a Lawfare piece on voluntary pre-deployment vetting, and, in this climate, "voluntary" is doing a lot of work. Here's Just Security:
He added that, as a result, “ffective immediately, no contractor, supplier, or partner that does business with the United States military may conduct any commercial activity with Anthropic.” Hegseth’s statement appears to indicate that this decision has been made and is final, although there is no information available that suggests the Department has effectuated it.
Hegseth posted on X that Anthropic is now a "Supply-Chain Risk to National Security," and that any contractor doing business with DoD has to cut commercial ties with Anthropic immediately. The catch is there’s still no evidence the Department has actually carried any of that out. He turned a social media post into a supposed policy move and called it corporate murder. But the authority to impose that kind of contractor blacklist does not just appear because the Secretary of Defense says it loud enough. Right. And Just Security is pretty clear on this: what Hegseth described as the consequences for Anthropic’s business partners does not line up with what he can actually order. Supply-chain risk designations have a process. A post is not that process. From Congressional Research Service:
On February 27, 2026, President Trump directed federal agencies to stop using technology developed by the U.S. artificial intelligence (AI) company Anthropic, and Secretary of Defense Pete Hegseth announced he was directing the Department of Defense (DOD) to designate Anthropic a "Supply-Chain Risk to National Security."
The Congressional Research Service just put out an In Focus brief on the federal government versus Anthropic, and it’s useful because CRS doesn’t have a side. It just lays out what happened, and what it means for procurement and competition. What happened is the Trump administration slapped Anthropic with a "supply-chain risk to national security" designation — the same label you use for a Chinese hardware vendor — in the middle of a contract fight. That’s not a security finding, that’s a cudgel. To be precise, the designation followed a reported months-long dispute between DOD and Anthropic over specific uses of its AI. CRS also flags that the effects go beyond Anthropic, because other domestic labs are watching how the "Secretary of War" handles AI vendor relationships. And every other lab is hearing the same message: push back on DOD use cases, and you could get labeled a national security risk by executive fiat. If you care about constraints on autonomous weapons, that is a nightmare incentive structure. Here's National Technology:
The US Department of Defense (DoD), also known as the Department of War, has signed agreements to deploy eight major AI companies’ technology on its private networks for “for lawful operational use”. The companies are SpaceX, OpenAI, Google, NVIDIA, Reflection.AI, Microsoft, Amazon Web Services, and Oracle.
Eight companies just got DoD network access at Impact Level six and seven, which means secret and above. And the phrase "lawful operational use" is doing a lot of heavy lifting in those agreements. I’d really like to see the contract language behind it. SpaceX, OpenAI, Google, Microsoft, AWS, Oracle, NVIDIA, and a company most people have never heard of, Reflection.AI, are all feeding into warfighter decision-making on the most restricted networks the Pentagon operates. Nobody’s defining where the AI recommendation ends and the human’s trigger finger begins. The DoD says 1.3 million personnel are already using GenAI.mil. That’s not a pilot program. That’s a deployment. These eight new deals just layer classified-network access on top of something that’s already baked into daily operations. And every one of those eight companies gets to put "DoD IL-6/7 certified" in the next enterprise sales deck. The national-security branding is worth nine figures before a single contract dollar even shows up. Lawfare writes:
The Trump administration is weighing the creation of a “review system” for frontier AI models. According to the New York Times, in this proposed approach, AI labs would provide the federal government with “first access” to “get ahead” of models with significant cyber capabilities, presumably such as Anthropic’s Mythos.
The Trump administration wants first access to frontier AI models before deployment, specifically the ones with significant cyber capabilities. There’s no clear legal mandate for that, so Lawfare is floating a voluntary framework: labs share with NIST’s CAISI, and CISA helps downstream actors patch whatever gets found. Voluntary. So the whole thing depends on labs deciding cooperation is good for PR. That’s not a safety regime, that’s a vibes agreement with a government logo on it. To be fair, Lawfare’s actual argument is that voluntary vetting could help labs avoid backlash for "knowingly introducing models that threaten critical systems" — which is doing a lot of legal work as a deterrent. The carrot is: do this now, or get something worse mandated later. And Anthropic gets name-checked specifically around Mythos and its cyber capabilities, so the subtext is pretty obvious: one lab is already being watched, and it would rather help shape the review process than get stuck with one it didn’t design. If Anthropic Pentagon Watch helps you stay on top of the story, please subscribe and leave a quick review wherever you’re listening. It really does help other people find the show.
We’ve put links to all of today’s stories in the show notes, so if one of them deserves a closer read, you can find it there.
That’s Anthropic Pentagon Watch for this Wednesday, May 6th. This is a Lantern Podcast.