Major financial backers of artificial intelligence company Anthropic are scrambling to resolve a bitter disagreement between the tech firm and military officials, according to seven sources with knowledge of the situation who fear the conflict could severely damage the company’s operations.
Chief Executive Dario Amodei has held conversations about the dispute with key investors and business partners in recent days, including Amazon.com’s CEO Andy Jassy, according to two individuals familiar with the discussions. Investment firms Lightspeed and Iconiq have also maintained contact with company leadership, sources revealed.
Several investors are reportedly contacting their connections within the Trump administration to help reduce tensions, two sources indicated.
The conversations center on preventing a complete prohibition of Anthropic’s artificial intelligence technology across all Pentagon contractor operations, according to those familiar with the matter.
Meanwhile, Anthropic and military officials continue limited discussions, though one source said Reuters could not confirm the specific nature of these conversations. President Donald Trump has demanded that Anthropic assist the government in eliminating its AI systems from federal use.
Amazon and the Pentagon did not provide immediate responses when contacted for comment.
The San Francisco-based AI developer and the Defense Department, recently renamed the Department of War by the current administration, have engaged in a prolonged disagreement spanning several months regarding military battlefield applications of the company’s technology. Industry observers view this confrontation as a crucial test of how much authority AI developers can maintain over their technological creations, which they believe could revolutionize educational systems, government services, and numerous other societal functions.
Military officials have urged AI companies to eliminate restrictions and instead agree to permit any lawful government use. However, Anthropic has maintained its position against allowing its Claude AI system to operate autonomous weapons systems or enable widespread domestic surveillance programs.
Among similar technology companies, Anthropic became the first to handle classified government information through a contract arrangement with cloud service provider Amazon. OpenAI announced Friday that it secured its own classified agreement with the Pentagon and stated that Anthropic should not be considered a security threat to the department.
During discussions with company executives, investors have confirmed their continued support for the San Francisco AI laboratory while simultaneously expressing their wish to reach an agreement with military officials, the seven sources said. Some financial backers told Reuters they felt frustrated that CEO Amodei had created antagonism with Pentagon leadership rather than building cooperative relationships. “It’s an ego and diplomacy problem,” explained one person briefed on the situation.
At this stage, some investors believe Amodei cannot appear to surrender to administration demands without alienating essential employees and customers who chose Anthropic specifically because of his principled position.
Amodei, who did not respond to requests for comment, has stated that Anthropic cannot “in good conscience accede to their request.” During a Tuesday evening conversation with investors, Amodei indicated the company would “continue to work to figure out a solution with the DoW.”
The investors intervening in Pentagon negotiations aim to help Anthropic avoid receiving a “supply-chain risk” classification from federal authorities, which could devastate the startup’s rapidly expanding commercial customer base.
Consumer interest has surged for Anthropic’s offerings, including its conversational AI Claude and programming tool Claude Code. On Monday, Claude ranked as the top free application download on Apple’s App Store, overtaking OpenAI’s ChatGPT.
Defense Secretary Pete Hegseth has indicated that such a risk classification would force all government contractors to cease using Anthropic’s technology throughout their entire operations. Anthropic has publicly disputed Hegseth’s statements, arguing he lacks legal authority to prevent use of its AI systems beyond defense-related contracts. The Pentagon did not respond to requests for comment regarding Anthropic’s assertion.
On Friday, Anthropic announced it would legally contest any supply-chain risk designation in federal court.
Nevertheless, some investors express concern that the confrontation might discourage potential clients who prefer to avoid any conflict with the current administration, one source noted.
These concerns arise during a pivotal period for the startup. Anthropic has secured tens of billions in funding based on ambitious projections for its business sales, which represent approximately 80% of the company’s income, according to company statements.
The outcome of future investment rounds, including a highly anticipated public stock offering, depends on Anthropic’s ability to continue expanding its commercial revenue. The company is currently allowing employees to sell shares to investors, though no final decision has been made regarding a public offering.
Anthropic’s revenue run rate, representing projected annual income based on current performance, has reached approximately $19 billion, one source revealed, increasing from $14 billion just weeks earlier.
The investor intervention occurred as multiple federal agencies began discontinuing their use of Anthropic’s technology, with the State Department switching to competitor OpenAI, following Trump’s Friday directive to eliminate Anthropic within six months.
