Trump Says White House Had "Very Good Talks" With Anthropic, Hints at Possible DOD Deal
President Trump has said his administration had "some very good talks" with Anthropic during a White House meeting last week and indicated that a deal between the US Department of Defense and the AI safety company is possible. The comments represent the first public acknowledgement of advanced discussions between the Trump administration and Anthropic regarding potential government AI deployments.
What Was Discussed
Details of the White House meeting remain limited, but Trump's comments suggest the discussions went beyond introductory conversations. The reference to a potential DOD deal is particularly significant — it implies Anthropic's Claude models may be under consideration for national security, intelligence, or military logistics applications. The NSA has previously been reported to be testing Anthropic models through a platform called Mythos, suggesting existing government-level engagement.
Anthropic's Government Strategy
Anthropic has been steadily building its government and enterprise presence. The company holds a Federal Risk and Authorization Management Program (FedRAMP) authorization, making Claude models eligible for federal agency use. Unlike some AI companies that have distanced themselves from defense contracts, Anthropic has not ruled out government deployments — framing responsible AI safety practices as compatible with national security use cases when appropriate safeguards are in place.
The AI-Government Race
The Trump administration has moved aggressively to integrate AI into government operations, signing executive orders to accelerate AI procurement and establishing a framework for evaluating frontier AI models for federal use. Anthropic's discussions follow OpenAI's own deepening relationship with the US government — OpenAI recently formed a national security subsidiary. The competition for major government AI contracts is intensifying.
Amazon's Role
Amazon's $25 billion investment in Anthropic adds a significant dimension to any government deal. AWS is already the primary cloud provider for the CIA and many other US intelligence agencies through its GovCloud infrastructure. A DOD deal with Anthropic would likely run on AWS, deepening Amazon's already substantial government cloud position.
The Bottom Line
Trump's public acknowledgement of Anthropic talks signals that the company is being seriously considered for major US government AI contracts. A DOD deal would be transformational for Anthropic's revenue, validation, and positioning — and would cement the Trump administration's preference for partnering with frontier AI labs rather than building government-owned AI systems.