Anthropic’s $200M Pentagon Deal Falls Apart — Now Negotiations Are Back On

TLDR

• A $200 million Pentagon agreement with Anthropic fell through when military officials insisted on removing contractual language that limited “analysis of bulk acquired data”
• Following the breakdown, President Trump ordered government agencies to cease using Anthropic’s products; Secretary Hegseth warned of potential national security supply-chain designation
• OpenAI swiftly secured its own Pentagon agreement, though CEO Sam Altman subsequently acknowledged rushing the process and advocated against penalizing Anthropic
• Public backlash propelled Claude to the top of Apple’s App Store, while Anthropic’s annualized revenue climbed from $14 billion to $19 billion
• CEO Dario Amodei has reengaged in discussions with Pentagon leadership to find common ground

---

Dario Amodei, chief executive of Anthropic, has restarted negotiations with Defense Department officials following last Friday’s breakdown in discussions regarding military deployment of Claude AI, The Financial Times reports.

Sources indicate Amodei is conducting meetings with Emil Michael, the Pentagon’s under-secretary for research and engineering, attempting to resurrect an agreement that would govern military access to Claude’s AI capabilities.

The initial contract, valued at $200 million, represented a watershed moment—making Claude the first major artificial intelligence system integrated into classified government infrastructure. The agreement was finalized by mid-2025.

Negotiations collapsed when Defense Department representatives requested that Anthropic remove specific contract language addressing “analysis of bulk acquired data.” In an internal communication, Amodei characterized this provision as protecting against his greatest concern.

Anthropic had pursued contractual protections preventing its technology from enabling domestic surveillance operations or autonomous weapons systems. Pentagon negotiators resisted these restrictions, maintaining that lawful military operations should determine technology deployment.

President Trump subsequently issued directives prohibiting federal agencies from utilizing Anthropic’s technology. Defense Secretary Pete Hegseth announced intentions to classify the company as a national security supply-chain vulnerability.

OpenAI Seizes Opportunity — Then Expresses Regret

Within hours of the White House’s action against Anthropic, OpenAI revealed it had secured its own Pentagon partnership. The announcement’s timing sparked immediate public criticism.

ChatGPT experienced a wave of application deletions, while Claude ascended to become Apple App Store’s leading download. Anthropic’s annualized revenue trajectory increased from $14 billion to $19 billion within weeks.

Key Anthropic investors, including Amazon’s CEO Andy Jassy, have engaged in direct conversations with Amodei regarding the situation. Additional investors contacted Trump administration officials seeking to prevent broader restrictions on Anthropic’s technology for defense contractors.

Certain investors expressed dissatisfaction, characterizing the dispute as substantially driven by “ego and diplomacy” rather than substantive policy disagreements.

Anthropic’s Financial Future Hangs in Balance

Enterprise customers generate approximately 80% of Anthropic’s total revenue. The company’s planned public offering relies substantially on maintaining robust revenue expansion.

An industry coalition including Nvidia, Google, and Anthropic submitted correspondence to Hegseth this Wednesday, voicing apprehension about designating an American technology company as a supply-chain security threat.

According to current reports, ongoing discussions between Anthropic and Pentagon officials continue, though specific terms remain undisclosed.

The post Anthropic’s $200M Pentagon Deal Falls Apart — Now Negotiations Are Back On appeared first on Blockonomi.