TLDR

  • The $200 million Pentagon agreement with Anthropic fell apart when the Department of Defense insisted on removing clauses that limited the “analysis of bulk acquired data.”
  • Federal agencies were ordered by President Trump to cease using Anthropic’s software, while Defense Secretary Hegseth warned that the company might be labeled a national security supply-chain risk.
  • Although OpenAI swiftly secured a Pentagon contract, CEO Sam Altman later admitted the company “shouldn’t have rushed” and advocated against classifying Anthropic as a supply-chain risk.
  • Following the public controversy, Anthropic’s Claude app reached the top of the Apple App Store, and annualized revenue surged from $14 billion to $19 billion.
  • Discussions have restarted between Anthropic CEO Dario Amodei and officials from the Pentagon.

According to The Financial Times, discussions regarding the military’s application of Claude AI collapsed on Friday, but Anthropic CEO Dario Amodei has resumed negotiations with the U.S. Department of Defense.

In a bid to save an agreement regarding the Pentagon’s access to Claude models, Amodei is reportedly holding discussions with Emil Michael, the under-secretary of defense for research and engineering.

Signed by the middle of 2025, the initial $200 million Pentagon contract established Claude as the premier significant AI model to be utilized within the government’s classified networks.

Negotiations stalled after the DoD requested the removal of a specific term regarding “analysis of bulk acquired data” from the agreement. In an internal memo, Amodei noted that this term addressed the precise scenario he was most concerned about.

Anthropic requested assurances that its software would not be employed for autonomous weaponry or domestic spying. The Pentagon countered, asserting its right to utilize the technology for any legal application.

Just days before the extensive collapse of discussions, Michael publicly criticized Amodei on X, labeling him a “liar” who possesses a “God complex.”

Subsequently, President Trump instructed federal agencies to discontinue the use of Anthropic’s tools. Defense Secretary Pete Hegseth stated his intention to classify the company as a national security supply-chain risk.

OpenAI Intervenes — Then Retreats

Mere hours after the White House’s action against Anthropic, OpenAI announced a new Pentagon agreement. This sequence of events sparked a public outcry.

While ChatGPT experienced a wave of uninstalls, Claude ascended to the number one spot on the Apple App Store. Over a few weeks, Anthropic’s annualized revenue rate increased from $14 billion to $19 billion.

OpenAI CEO Sam Altman subsequently remarked that his firm “shouldn’t have rushed” into the agreement. Additionally, he stated on X that Anthropic ought not to be labeled a supply-chain risk and that the DoD should extend to Anthropic the same conditions accepted by OpenAI.

Investors in Anthropic, such as Amazon CEO Andy Jassy, have spoken directly with Amodei regarding the circumstances. To prevent a wider prohibition on Anthropic’s AI for all Pentagon contractors, other investors contacted the Trump administration.

Certain investors expressed frustration, attributing the issue to “ego and diplomacy” as much as to the actual policy details.

Anthropic’s Business on the Line

Approximately 80% of Anthropic’s revenue is derived from enterprise sales. The company’s projected IPO relies significantly on sustained revenue expansion.

On Wednesday, a technology industry coalition comprising Nvidia, Google, and Anthropic sent a letter to Hegseth voicing concerns over classifying a U.S. firm as a supply-chain risk.

Supporters assembled outside Anthropic’s headquarters, inscribing messages in chalk on the sidewalk. One message declared: “God loves Anthropic.”

According to the most recent updates, discussions between Anthropic and the Pentagon are ongoing, although the specific terms are still uncertain.