Few situations in the modern technology landscape illustrate regulatory contradiction more starkly than the position Anthropic currently occupies: its artificial intelligence systems are being actively deployed in an ongoing military conflict, even as defense contractors race to remove those same systems from their operations. The company finds itself at the center of a geopolitical and regulatory storm that has no clear precedent in the AI industry.
The origins of this contradiction lie in a series of overlapping and conflicting directives issued at the highest levels of the U.S. government. President Trump instructed civilian agencies to discontinue the use of Anthropic products, while simultaneously granting the company a six-month window to wind down its relationship with the Department of Defense. Before that directive could be meaningfully executed, the United States and Israel launched a surprise military strike on Tehran, dragging Anthropic's technology into an active theater of war.
The operational implications of this timing are significant. Because no formal legal barriers have yet been established — and because Secretary of Defense Pete Hegseth has not yet taken official steps to designate Anthropic as a supply-chain risk despite publicly pledging to do so — there is nothing preventing the continued use of Anthropic's models in military decision-making. The gap between political intent and regulatory action has created a window in which the technology remains legally permissible, regardless of the controversy surrounding it.
A Washington Post investigation published Wednesday provided some of the most detailed public accounting yet of how Anthropic's systems are functioning within the current conflict. Working in conjunction with Palantir's Maven system, the AI platforms reportedly
"suggested hundreds of targets, issued precise location coordinates, and prioritized those targets according to importance,"with the Post characterizing the overall function as "real-time targeting and target prioritization." The disclosure adds considerable weight to the debate over AI accountability in combat environments.
On the commercial defense side, the response has been swift and decisive. Lockheed Martin and other major defense contractors moved this week to replace Anthropic's models with competing alternatives, according to a Reuters report. The displacement extends well beyond prime contractors. A managing partner at J2 Ventures told CNBC that 10 of his portfolio companies
"have backed off of their use of Claude for defense use cases and are in active processes to replace the service with another one."The scale of this withdrawal signals a broader recalibration across the defense technology ecosystem.
The central unresolved question now is whether Hegseth will formalize the supply-chain risk designation — a move that would almost certainly trigger significant legal proceedings. Until that determination is made, Anthropic occupies a deeply paradoxical position: being systematically removed from the defense industry's commercial infrastructure while simultaneously operating within one of the most sensitive military engagements of the current era. How that contradiction resolves will likely shape the governance of AI in national security contexts for years to come.
