Trump Bans US Agencies From Using Anthropic’s AI, Sparks Tech Industry Outcry

Washington D.C. – The United States government’s dispute with artificial intelligence firm Anthropic escalated sharply on Friday, as former President Donald Trump directed all federal agencies to cease working with the company’s AI technologies, effective immediately. Trump announced the directive on his Truth Social platform, stating, “We don’t need them, we don’t want them, and we will not be doing business with them anymore.” He further characterized Anthropic as “leftwing nut jobs” and “woke.”

While most agencies are to immediately halt engagement, the Pentagon has been granted a six-month period to phase out the technology, according to Trump’s statement. He also threatened the company with “civil and criminal consequences” should they not be “helpful” during the transition. This move signals a significant hardening of the administration’s stance towards Anthropic, a leading developer of advanced AI models.

The escalation came as Defense Minister Pete Hegseth went even further, designating Anthropic as a “supply chain risk to national security.” This classification, rarely applied to companies outside of those with ties to China or Russia, effectively places Anthropic on a blacklist, restricting its access to government contracts and partnerships.

The implications for Anthropic are substantial. Hegseth indicated that Pentagon contractors would be prohibited from conducting any business with the company. Anthropic maintains close working relationships with major American technology firms, including Nvidia, Google, Amazon, and Microsoft, with some of these companies holding equity stakes in the AI developer. Disrupting these partnerships could have a cascading effect on the broader tech landscape.

Dean Ball, an AI policy analyst and former technology advisor to the White House, described the government’s actions as a “attempted corporate murder” and a “psychotic grab for power,” suggesting the move is “almost certainly illegal.” He argued that the administration is sending a message that This proves an unreliable partner for businesses.

Ultimatum Lapsed Friday

Hegseth had issued an ultimatum to Anthropic earlier this week, which expired late Friday afternoon. The ultimatum threatened sanctions if the company did not agree to loosen existing restrictions on the use of its AI models. Beyond the revocation of a $200 million contract and the designation as a supply chain risk, Hegseth also indicated the possibility of compelling Anthropic to alter its terms of service through the invocation of the “Defense Production Act,” a law typically reserved for national emergencies like the COVID-19 pandemic. Hegseth chose to end business relationships with Anthropic and declare it a risk to national security.

In a post on X, Hegseth stated, “Anthropic’s posture is fundamentally incompatible with American principles.” He asserted that the relationship between the company and the government had been “permanently” altered, adding, “This decision is final.”

The core of the dispute centers on two limitations Anthropic places on the use of its AI models: they are not to be used for mass surveillance of Americans or for the development of fully autonomous weapon systems. The Pentagon, however, insists on the right to utilize the technologies for “all lawful purposes.”

Amodei: “Incompatible with Democratic Values”

Anthropic CEO Dario Amodei remained resolute on Thursday, writing in a lengthy blog post that the company could not, in good conscience, comply with the Pentagon’s demands. He explained that deploying the technology for domestic mass surveillance is “incompatible with democratic values,” and that AI models are currently “not reliable enough” for autonomous weapon systems.

Amodei labeled the Pentagon’s threats – to designate Anthropic a supply chain risk and invoke the Defense Production Act – as “contradictory.” He argued that the company was being simultaneously portrayed as a security risk and as “essential” to national security. Amodei expressed hope that the government would reconsider its position, stating that if not, Anthropic would assist in “a smooth transition to another provider.” He emphasized his belief in the “existential importance” of utilizing AI for the defense of the United States and other democracies.

Trump characterized Anthropic’s attempt to impose its terms on the Pentagon as a “disastrous mistake.” “The USA will never allow a radical left, woke company to dictate how our great military fights and wins wars,” he wrote. He added that Anthropic is “run by people who have no idea what the real world is about.”

Anthropic received unexpected support from its rival, OpenAI. Sam Altman, OpenAI’s CEO, communicated to employees – as reported by the Wall Street Journal – that his company maintains similar restrictions. He stated he would push for the exclusion of areas like domestic surveillance and autonomous weapons from government contracts.

Altman and Amodei have a strained relationship, reportedly refusing a symbolic handshake at a recent AI summit in India.

Solidarity with Anthropic in Silicon Valley

Employees from OpenAI and Google expressed solidarity with Anthropic, appealing to their companies’ leadership to maintain a firm stance and not concede to the Pentagon’s demands. As of Friday, the letter had been signed by more than 500 Google employees and over 80 OpenAI staff members.

Separately, over 100 Google employees sent a letter to company leadership, urging them to uphold the same restrictions as Anthropic. Employee groups from Google, Amazon, and Microsoft also issued a joint open letter, calling on their companies to reject the Pentagon’s demands.

Anthropic’s Claude AI model is considered among the best in the industry and is reportedly heavily utilized by the government. Sources within the Pentagon suggest that abandoning Anthropic would represent a significant loss.

In the past year, OpenAI, Google, and X.AI, led by Elon Musk, have also received $200 million in contracts from the Pentagon. Until recently, Anthropic was the sole partner whose AI models were cleared for handling classified information.

Recently, it was revealed that X.AI has also been granted access to this level of classified work. However, it remains unclear whether X.AI’s Grok model could replace Claude. The Wall Street Journal reported that representatives from several government agencies have expressed concerns about the safety and reliability of Grok in recent months.

X.AI would undoubtedly be a potential beneficiary in the current dispute between Anthropic and the U.S. Government. Musk has repeatedly attacked Anthropic in recent times, writing on his platform X on Friday: “Anthropic hates Western civilization.”