Anthropic Files Two Federal Lawsuits Against the Trump Administration
Anthropic filed two federal lawsuits today against the Trump administration, calling the Pentagon's actions an "unlawful campaign of retaliation" against a private company for holding protected First Amendment views on AI safety. One 48-page complaint, filed in the U.S. District Court for the Northern District of California, argues the government violated the Constitution and exceeded its statutory authority. The second, filed in the U.S. Court of Appeals for the D.C. Circuit, directly targets the "supply chain risk" designation and asks the court to reverse it — a route available because federal law grants that court direct jurisdiction over such findings. The White House responded that the administration will not allow warfighters to be "held hostage by the ideological whims of any Big Tech leaders."
How a Biden-Era Contract Became a Constitutional Standoff
The conflict traces to a defense contract negotiated under Biden containing two hard limits: Claude could not be used for mass domestic surveillance of Americans, nor to control fully autonomous lethal weapons. The Trump administration initially accepted those terms, then demanded renegotiation to allow Claude to be used for "all lawful purposes."
Talks collapsed on February 27, when Anthropic's CEO Dario Amodei personally told Defense Secretary Pete Hegseth the guardrails were permanent. The Pentagon's chief technology officer argued the limits were redundant — surveillance is already illegal, autonomous weapons already restricted internally. Amodei held firm: legal prohibitions without explicit contractual guarantees were insufficient. Trump ordered every federal agency to immediately cut ties with Anthropic. On March 4, Hegseth formalized the break with the supply-chain-risk designation — a label ordinarily reserved for foreign adversaries like Chinese telecoms — directing that no military contractor "may perform any commercial activities with Anthropic."
The Legal Architecture: Strong in Theory, Uncertain in Court
Anthropic's case rests on three arguments: the statutory definition of supply-chain risk targets foreign sabotage, not a domestic company's contract terms; punishing a company for publicly articulating safety positions constitutes First Amendment retaliation; and no federal statute authorizes the president to sever the entire government from a private vendor by executive order. Legal analysts broadly view the domestic application of the designation as procedurally vulnerable.
The government's strongest counter is judicial deference on national security. The likely near-term outcome is partial relief on scope or process — while the administration reroutes spending through other channels regardless.
The Battlefield Irony Every Investor Should Know
The most corrosive detail for the government's credibility: the U.S. military reportedly continued using Claude during operations against Iran that began in late February, even after the government-wide ban. The Pentagon relies on Claude primarily for document synthesis and logistics efficiency. Replacing that capability could take three months or longer — a switching cost that makes a quiet, face-saving settlement commercially plausible.
Multiple Compression, Not Revenue Loss
The critical mistake is treating this as a revenue story. Anthropic closed a recent funding round at a reported $380 billion post-money valuation, with annualized revenue approaching $7 billion. Direct defense contracts are not the valuation driver — contagion is. The designation forces contracting lawyers at every defense vendor, regulated bank, and Fortune 500 procurement team to open a review. Procurement teams overcorrect under uncertainty: freezes, architectural rewrites, rival pilots. Political fragility discounts can dwarf actual contract losses for a company priced on hypergrowth and optionality.
The clearest beneficiary is OpenAI, which has already secured a Pentagon deal on more flexible terms. Second-order winners include defense integrators — Palantir and Amazon-mediated environments were already the primary channels for Claude's defense deployments — and Google and xAI, both deepening their Pentagon presence.
The strategic fork is now explicit: Anthropic monetizes trust; OpenAI monetizes ubiquity. In defense procurement, ubiquity wins. Both models survive at scale — but not inside the same contracts.
The deepest question this case forces is not legal. It is who controls the last-mile use policy of general-purpose AI — the lab, the customer, or the state. That is now a first-order valuation variable for every frontier AI investment on earth.
not investment advice
Sources: Case 1 — U.S. District Court, Northern District of California https://cand.uscourts.gov/cases-e-filing/cases/
Case 2 — U.S. Court of Appeals, D.C. Circuit https://www.cadc.uscourts.gov/internet/home.nsf