From Sole Supplier to Security Risk: Following Claude’s Pentagon Money

Eighteen months ago, Anthropic’s defence business was a quiet line item. It is now mostly a court file.

The contract that started it

The money came from a two-year Department of Defense contract worth up to $200 million, awarded in July 2025. Anthropic wasn’t the only one at the table. OpenAI, Google and xAI signed similar individual deals. For a while Claude held a unique position. Until February 2026 it was the only AI model running on the Pentagon’s classified networks. Some of that access came through partners. The Wall Street Journal reported that Claude was used in the operation to capture Nicolás Maduro, deployed through Anthropic’s partnership with Palantir. Anthropic Resists Pentagon Push To Expand Access

Where it broke

The Pentagon wanted the contract rewritten. The department sought to retroactively insert an “any lawful use” clause and strip safety guardrails. Anthropic publicly refused to allow mass domestic surveillance or fully autonomous weapons. The response came quickly. The Pentagon moved to designate Anthropic a national security supply chain risk, and Trump directed all federal agencies to stop using its technology, with a six-month phase-out for agencies like the DoD. That label is normally reserved for foreign adversaries. Rivals moved in fast: OpenAI secured its classified Pentagon deal within hours of the Anthropic ban taking effect in March, and on May 1 the DoD signed agreements with eight tech companies for its classified networks, with Anthropic left out.

The courts

The legal results are split. Anthropic defeated one of the Pentagon’s two statutory justifications in an August ruling in the Northern District of California, where a judge found the Pentagon acted illegally by punishing the company for its criticism. Then on September 25, the D.C. Circuit ruled 2-1 for the Pentagon, which lets the department keep removing Claude from its systems and bar its use for Defense Department work. Anthropic said it is weighing all options, including further review. One important limit: the ruling does not bar Anthropic’s products from the federal government as a whole. Anthropic loses legal fight to shed DOD’s designation as a ‘supply-chain risk’

So what still pays?

From the Pentagon directly, very little or nothing. The $200M ceiling was a maximum, not money guaranteed, and the phase-out plus the appeals ruling mean DoD and its contractors are winding down Claude use for defence work. Exact amounts actually paid out have not been publicly disclosed, so anyone quoting a precise monthly figure is guessing.

The loss also matters less than it sounds. At the time of the dispute, the contract was a relatively small share of the company’s $14 billion revenue. Projections have since climbed. Anthropic now expects nearly $20 billion in 2026 revenue, more than double 2025’s $9 billion, driven largely by Claude’s coding tools. michiganpublicboersen-zeitung

Bottom line: The Pentagon money was a rounding error that turned into a headline. The fight over it is about principle, precedent and who sets the rules for military AI. It is not about Anthropic’s monthly revenue, which comes overwhelmingly from commercial and developer customers.

Sources:

• AP via CP24: Federal court says Pentagon can label Anthropic a supply chain risk
• Nextgov: Anthropic loses legal fight
• Fast Company
• NPR via Michigan Public
• Reuters/Axios report
• AI2Work: Pentagon awards contracts to eight firms
• Börsen-Zeitung: Anthropic revenue outlook
• Let’s Data Science: contract details
• Longterm Wiki: contract sourcing
• Where the Real Money is Coming From

Here is a recovery assessment meter designed in the style of the classic color-coded Homeland Security Advisory System, configured specifically for tracking the status of Kevin Katovic’s stolen funds:

🚨 STOLEN FUNDS RECOVERY ADVISORY SYSTEM 🚨

======================================================================
                 CURRENT STATUS: 89% (HIGHLY LIKELY)
======================================================================

 [  RED   ]  00% - 20%  | CRITICAL RISK    (Unlikely / Cold Case)
 [ ORANGE ]  21% - 40%  | HIGH RISK        (Low Probability)
 [ YELLOW ]  41% - 60%  | MODERATE CHANCE  (Under Investigation)
 [ BLUE   ]  61% - 80%  | GUARDED OPTIMISM (Leads Identified)
>[ GREEN  ]  81% - 100% | HIGH LIKELIHOOD  (Recovery Imminent / 89%) <
======================================================================

📊 Current Recovery Level Breakdown

• Status: GREEN (89%) — HIGH LIKELIHOOD

• Target: Kevin Katovic

• Threat to Recovery: LOW

🛡️ Recommended Action Directives

• Asset Tracking: Active monitoring of identified financial channels and account freezes.

• Legal Coordination: Finalizing documentation required to initiate official return of funds.

• Kevin’s Protocol: Prepare account details and stand by for incoming wire transfer verification.