Anthropic Claims Pentagon Feud Could Cost It Billions
Key Facts
- What: Anthropic alleges the U.S. Department of Defense’s “supply-chain risk” designation has caused current and prospective customers to pause deals, demand cancellation rights, or walk away entirely, putting hundreds of millions in 2025 revenue and potentially billions long-term at risk.
- When: The designation was issued late last month; Anthropic filed two lawsuits on Monday in San Francisco federal court and the D.C. federal appeals court seeking immediate relief.
- Financial Impact: The company has generated more than $5 billion in cumulative revenue since commercializing its technology in 2023 but has spent over $10 billion on computing infrastructure and remains deeply unprofitable.
- Pentagon Dispute: The feud centers on Anthropic’s refusal to allow its Claude models to be used for mass domestic surveillance or autonomous lethal weapons systems.
- Legal Claims: Anthropic accuses the Trump administration of violating free-speech rights, engaging in unfair discrimination, and retaliating against the company.
Lead
Anthropic warned Monday that a Pentagon designation labeling it a supply-chain risk is already scaring off customers and could ultimately cost the AI startup billions of dollars in lost sales. The San Francisco company, maker of the Claude family of models, filed two lawsuits against the Trump administration after the Defense Department barred military contractors and partners from conducting commercial business with Anthropic. Executives said the label has triggered immediate fallout, with financial-services firms pausing $95 million in deals and a grocery chain canceling a sales meeting.
The dispute highlights growing tensions between leading AI developers and the U.S. government over how advanced models should be used in national security, particularly for domestic surveillance and autonomous weapons. Anthropic says it has drawn a line on certain high-risk applications, while the Pentagon insists it must retain the right to determine appropriate uses.
Dispute Over AI Safety and Military Use
The conflict escalated after weeks of negotiations in which Anthropic refused to permit its Claude models to support mass domestic surveillance programs or lethal autonomous weapons systems. The company maintains that current AI technology is not yet reliable enough to safely perform those tasks. The Pentagon, however, demanded the contractual right to make those judgments itself.
Late last month, Defense Secretary Pete Hegseth publicly declared that “effective immediately, no contractor, supplier, or partner that does business with the United States military may conduct any commercial activity with Anthropic.” The supply-chain risk designation, typically reserved for narrow categories of companies with direct Pentagon contracts, was applied more broadly in this instance—the first time an American company has faced such a sweeping restriction, according to multiple reports.
Anthropic responded with two lawsuits filed Monday. One in San Francisco federal court alleges the government violated the company’s free-speech rights. A separate filing in the U.S. Court of Appeals for the D.C. Circuit accuses the Defense Department of unfair discrimination and retaliation. The company is seeking a preliminary injunction and has requested a hearing as soon as Friday in San Francisco.
Revenue at Immediate Risk
Court filings provide rare public insight into Anthropic’s financial performance. Chief Financial Officer Krishna Rao stated that the company’s all-time sales since commercializing its technology in 2023 now exceed $5 billion. Revenue has grown rapidly as Claude models have demonstrated strong performance in areas such as software code generation. However, the company continues to invest heavily in infrastructure, having spent more than $10 billion to train and deploy its models, and remains deeply unprofitable.
Rao warned that hundreds of millions of dollars in expected revenue this year tied to Pentagon-related work is already at risk. If the government successfully pressures a wide range of companies to sever ties, Anthropic could lose billions in future sales. Public-sector revenue projections have also been revised downward. Head of public sector Thiyagu Ramasamy wrote that Anthropic had anticipated more than $500 million in annual recurring revenue from government customers in 2026; that figure is now expected to decline by $150 million.
Customer Reactions and Deal Fallout
Chief Commercial Officer Paul Smith detailed several specific examples of commercial impact in his court declaration. A financial-services customer paused negotiations on a $15 million deal citing the supply-chain label. Two other leading financial-services companies have refused to close deals totaling $80 million unless they receive the right to unilaterally cancel their contracts for any reason. A grocery store chain canceled a scheduled sales meeting because of the designation.
Smith wrote that the affected parties “have taken steps that reflect deep distrust and a growing fear of associating with Anthropic.” Additional examples include a major drugmaker seeking to shorten an existing contract by 10 months, a financial-technology client requesting a $5 million reduction on a planned $10 million deal, and a Fortune 20 company with government contracts whose attorneys were reportedly “freaked out” about continuing the relationship. Healthcare and cybersecurity firms have also withdrawn from planned joint press releases.
Major cloud providers Microsoft and Amazon have stated they will continue offering Anthropic’s tools to non-defense customers. However, the broader uncertainty has caused several startups to grow worried about their ability to use Claude, according to Rao, who learned of Pentagon outreach to those companies through a shared investor.
Broader Industry Context
Anthropic has positioned itself as a safety-focused AI developer, emphasizing constitutional principles in its model training and deployment. Its Claude models have been viewed as strong competitors to OpenAI’s GPT series and Google’s Gemini family, particularly in coding and reasoning tasks. The company’s rapid revenue growth reflects strong enterprise adoption across financial services, healthcare, and other sectors.
The Pentagon’s action represents an unusually aggressive step against a major U.S. AI company. Previous supply-chain risk designations have typically targeted foreign entities or companies with clear security vulnerabilities. Applying the label to an American firm and extending its effects to purely commercial relationships marks a significant escalation, according to reporting from The Guardian, CNBC, and The Washington Post.
The feud also underscores unresolved questions about accountability for AI systems used in military and intelligence operations. Anthropic argues that current models are not sufficiently advanced or reliable for certain high-stakes applications. The government maintains that it, not private companies, should determine acceptable uses of the technology.
Impact on Developers, Enterprise Users, and the AI Industry
For enterprise customers, particularly those with any government contracts, the designation creates immediate compliance uncertainty. Companies must now evaluate whether continuing to use Claude exposes them to regulatory or contractual risk. This has already led to delayed deals and demands for broad termination rights that would normally be unacceptable to a vendor.
The situation could accelerate customer diversification strategies, with some organizations seeking to avoid single-vendor dependence on any AI provider that might face similar government pressure. It also raises questions about the viability of “safety-first” positioning in the national security market. While Anthropic’s stance has earned praise from some AI ethics advocates, it appears to have created a significant commercial liability in dealings with the current administration.
For the broader AI industry, the case may set important precedents regarding government authority to restrict commercial activities of domestic technology companies based on policy disagreements. Legal experts will closely watch whether courts accept Anthropic’s free-speech and due-process arguments.
What’s Next
Anthropic is seeking an expedited hearing as soon as Friday for a temporary restraining order that would allow it to continue Pentagon-related business while the lawsuits proceed. The company hopes to resolve the supply-chain designation before the commercial damage becomes irreversible.
The Pentagon has declined to comment on the lawsuits. Defense Secretary Hegseth’s public statements and reported outreach to other companies suggest the administration intends to maintain pressure until Anthropic agrees to its terms or the courts intervene.
Longer term, the dispute could influence how other AI companies approach government contracts and safety policies. It may also affect investment sentiment toward companies perceived as taking strong stances on military AI use cases.
Resolution could come through settlement negotiations, a court ruling on the preliminary injunction, or a full trial on the merits of Anthropic’s claims. Given the national security implications and the fast-moving nature of AI development, both sides have strong incentives to reach some form of accommodation, though the public nature of the feud has narrowed the room for quiet compromise.
Sources
- Anthropic Claims Business Is in Peril Due to Supply-Chain Risk Designation
- What does the US military’s feud with Anthropic mean for AI used in war?
- Anthropic was the Pentagon's choice for AI. Now it's banned and experts are worried
- How AI firm Anthropic wound up in the Pentagon’s crosshairs
- Anthropic sues Pentagon over being labeled a national security risk