Trump Cuts Ties with AI Firm Anthropic Over Security Concerns & Control Dispute

WASHINGTON — President Donald Trump ordered U.S. Government agencies to cease all use of technology from Anthropic, an artificial intelligence company, following a weekslong standoff over the military’s desired applications of its AI models. Trump announced the decision Friday afternoon on Truth Social, stating, “We don’t demand it, we don’t want it, and will not do business with them again!”

The General Services Administration (GSA) swiftly moved to implement the order, suspending Anthropic’s access through USAi, a platform providing chatbots to government workers, according to an email obtained by this publication. The government is likewise initiating the removal of Anthropic from its primary procurement system, effectively barring federal agencies from purchasing its products.

Anthropic, which last summer secured a $200 million Pentagon contract, had distinguished itself among AI firms like OpenAI, Google, and xAI by being the only one to receive security clearance permitting the handling of classified data with its Claude models. These models were reportedly integrated across the Department of Defense, and, according to sources, assisted in operations such as the capture of Venezuelan President Nicolás Maduro.

The dispute centers on Anthropic’s refusal to allow Claude to be used for mass domestic surveillance or to power fully autonomous weapons systems – systems capable of selecting and engaging targets without human intervention. Anthropic maintains that these uses were never part of its contracts with the Pentagon. However, the Department of Defense (DOD) now demands unrestricted access to Claude, accusing Anthropic of hindering national security. As reported by CNBC, the Trump administration views Anthropic’s stance as an attempt to control the military.

Following a Tuesday meeting, the DOD gave Anthropic until 5:01 p.m. Eastern time Friday to comply. Failure to do so, the Pentagon warned, would result in the invocation of the Defense Production Act or, more severely, a designation as a national security “supply-chain risk,” effectively barring any company working with the U.S. Military from doing business with Anthropic. Shortly after Trump’s announcement, Defense Secretary Pete Hegseth followed through on the threat, declaring Anthropic a supply-chain risk, as the Associated Press reported.

Anthropic CEO Dario Amodei publicly stated his company’s unwillingness to concede. In a statement released Friday, Anthropic affirmed, “No amount of intimidation or punishment from the Department of War will change our position.” The DOD, which the Trump administration now refers to as the Department of War, did not respond to requests for comment.

The situation highlights a growing tension between Silicon Valley and the federal government. Concerns are mounting that the U.S. Military is falling behind adversaries, particularly China, in technological advancements, as much innovation now originates in the private sector. As The New York Times noted, the massive investments required for generative AI are largely coming from tech companies, not federal grants.

Historically, the Pentagon has not dictated terms of use for purchased technologies. However, Anthropic’s security clearance and its insistence on ethical boundaries have upended this dynamic. Amodei, despite being a proponent of using AI to counter authoritarian regimes, has expressed reservations about deploying AI-powered weapons before the technology is sufficiently reliable, fearing risks to both soldiers and civilians.

The White House has reportedly been increasingly critical of Anthropic in recent months, with officials labeling the company “woke” and accusing it of “fear mongering,” following Amodei’s public criticism of Trump and support for Kamala Harris during the 2024 election. This has created a paradoxical situation, as the administration simultaneously deems Claude essential for national security while questioning its trustworthiness.

The fallout from the supply-chain-risk designation remains uncertain. It could potentially force major tech companies like Google, Microsoft, and Amazon – all federal contractors – to sever ties with Anthropic, a move that could be devastating for the AI firm, particularly given Amazon’s investment in data centers to train future versions of Claude. However, the extent of the impact is still being debated.

Meanwhile, OpenAI CEO Sam Altman circulated an internal memo, obtained by this publication, stating that his company “has long believed that AI should not be used for mass surveillance or autonomous lethal weapons.” More than 500 employees of OpenAI and Google signed an open letter supporting Anthropic. Despite this show of support from within Silicon Valley, the future of the relationship between the U.S. Government and Anthropic remains unclear, as the Pentagon has six months to phase out Claude, suggesting its continued importance despite the current conflict.

Altman has indicated that OpenAI is exploring a contract with the Pentagon for classified workloads, contingent on excluding uses deemed unlawful or unsuitable for cloud deployments. The Pentagon reportedly agreed to these conditions, though no contract has been signed. The showdown between the Pentagon and Anthropic may foreshadow a broader conflict within Silicon Valley itself, as differing views on the ethical and strategic use of AI emerge.