President Donald Trump said Sunday, Sept. 27, that he planned a White House dinner with Anthropic CEO Dario Amodei, following months of conflict over the company’s limits on military uses of its AI. The plan came two days after a federal appeals court upheld a Pentagon supply-chain exclusion of Anthropic under one procurement law. Public reporting has not independently confirmed that the dinner took place or revealed what the two discussed, Euronews reported.
Trump announced the plan publicly. “And tonight, I'm having dinner with the head of Anthropic, so these are long days, but I seem to be able to handle it. Right” he said, as reported by Euronews. That announcement establishes what he intended, not whether the meeting happened or who else may have been there.
The distinction matters because the dispute concerns who sets limits on AI used by the government. A dinner between a president and a technology executive does not establish a policy bargain. But any private discussion about federal AI procurement or military use would concern decisions with consequences for civil liberties and public accountability.
What are Trump and Amodei in conflict over?
The disagreement centers on the Pentagon’s request for broad authorization to use Anthropic’s AI for any lawful purpose, and the company’s refusal to remove two safeguards. Anthropic has said it would not allow its models to be used for mass domestic surveillance or fully autonomous weapons.
The company has argued that domestic surveillance raises risks to democratic liberties and that fully autonomous weapons raise reliability concerns. It has also said it wanted to continue working with the Department of War while keeping those limits in place, and offered to collaborate on research to improve system reliability.
Those are Anthropic’s stated concerns, not evidence that the government used Claude for either prohibited purpose. The central dispute is over who should set the boundaries: the company through contract terms, or the department through its demand for broad authorization.
The disagreement followed a period of government use of Claude. Anthropic’s account says the system had been deployed across the Department of War and other national-security agencies for tasks including intelligence analysis, operational planning and cyber operations. The legal record describes early use as limited and routed through contractors. The available reporting does not establish Claude’s current status across government systems.
What did the appeals court decide?
The court’s account traces the dispute to February. Defense Secretary Pete Hegseth met Amodei on Feb. 24 and sought acceptance of the contract term by Feb. 27. Amodei made Anthropic’s refusal public on Feb. 26. The secretary announced an intent to exclude the company from the department’s supply chain the next day, with a formal written decision on March 3.
On Sept. 25, the D.C. Circuit denied Anthropic’s petitions challenging that exclusion under the procurement law before it, finding the department’s action lawful. One judge dissented. The court also noted that a district court had set aside a separate designation under a different statute on Aug. 27. The appeals ruling therefore did not resolve every legal issue involving Anthropic’s government status.
The court found that Anthropic’s advocacy for safeguards was protected speech and that the exclusion materially harmed the company. But it found no causal link between that advocacy and the department’s action, concluding the department acted over Anthropic’s refusal to accept the contract term.
Free newsletter
Get the morning briefing
Start each day with the stories that matter and why — a short, free email from our newsroom.
The ruling did not establish that Anthropic’s safeguards were unlawful or prove either side’s forecast about possible AI risks. The court described competing concerns about a constrained model failing during an operation and an unconstrained model producing a false target. Those were arguments in the dispute, not findings that either outcome had occurred.
What remains private about the dinner?
Trump has argued that major AI companies should not slow development because China could gain an advantage. Euronews also reported that he dismissed concerns about security incidents involving advanced AI, saying, “I'm not worried about it” The sources reviewed do not establish whether that disagreement, the court ruling or any other issue came up at the planned dinner.
That gap is a reason for scrutiny, not a basis to assume an agreement. Federal decisions about AI procurement and military use affect how public agencies deploy powerful systems and what safeguards apply. The public can assess those choices only if the government explains its decisions and any commitments made with companies seeking or holding public contracts.
Three questions remain open:
- Did the planned dinner take place, and what did Trump and Amodei discuss?
- Who else, if anyone, attended?
- Did the meeting produce any commitments or formal follow-up, and will those be made public?