OpenAI and Google researchers file amicus brief backing Anthropic in Pentagon case

• More than 30 OpenAI and Google workers, including Google’s Jeff Dean, have filed an amicus brief backing Anthropic in its fight with the Pentagon, according to the brief report.
• The dispute centers on a Pentagon decision to label Anthropic a “supply chain risk,” which Anthropic argues is an extraordinary designation for a U.S. company rather than a foreign adversary, as described in the case summary and the BBC summary.
• Anthropic is seeking relief in two courts to remove the blacklist label and stop agencies from cutting it off, with the filing recap saying the company is also framing the move as retaliation tied to its advocacy on AI safety rules for weapons.
• For engineers working on federal deployments, the practical stake is procurement access: one post says the order tells agencies to “rip out” Anthropic systems from operations, as shown in the quoted order.

Anthropic has now turned its Pentagon blacklist fight into a broader industry test. According to the BBC summary, the company filed lawsuits in two separate courts to get the “supply chain risk” label removed and to stop agencies from cutting off access to its products.

The public summaries describe Anthropic’s argument in two parts. First, the case summary says Anthropic claims the designation is a “gross overreach” because the label is normally used for hostile foreign actors, not a domestic AI vendor. Second, the filing recap says Anthropic is arguing the government used the designation to punish the company over its speech and policy positions on AI safety for weapons.

The new detail is not just that Anthropic is suing; it is that rivals’ employees are lining up behind the challenge. The amicus report says more than 30 experts from OpenAI and Google joined the brief, and names Jeff Dean among the signers. The same recap says those supporters argue the blacklisting could damage U.S. AI leadership, which turns the case from a single-vendor dispute into a precedent fight over how federal agencies can exclude model providers.

The operational consequence appears immediate. One widely shared quote from the order says the government was told to “rip out” Anthropic AI from its operations, according to the quoted order. A separate post also shows outside groups are already organizing additional amicus support around the case via a support call, suggesting this will be watched as a procurement and platform-access dispute, not just a speech fight.