Pentagon Labels AI Firm Anthropic a ‘Supply Chain Risk’ in High-Stakes Legal Battle

By Jim Thomas | Thursday, March 19, 2026

Anthropic co-founder Jack Clark met privately with members of the House Homeland Security Committee on Wednesday for a bipartisan briefing on artificial intelligence security issues, including how advanced systems can be copied and how sensitive technology is shared across borders.

Clark, who recently took on a new role as head of public benefit, met with lawmakers from both parties in a closed-door session.

The meeting comes as Anthropic fights a high-stakes court battle over a Pentagon designation labeling it a “supply chain risk,” following a clash with the Trump administration over how the military can use the company’s AI systems. The dispute was only briefly discussed, with lawmakers focusing instead on national security, cybersecurity, and AI governance.

Anthropic sued on March 9 after the Pentagon formally designated the San Francisco tech company a supply chain risk following an unusually public dispute over how its AI chatbot Claude could be used in warfare. The company stated that the action followed its refusal to remove safeguards against uses such as autonomous weapons and domestic surveillance, while the administration claimed the dispute arose from procurement and national security concerns.

The legal fight has intensified in recent days. Anthropic has asked a federal appeals court to pause the designation, arguing it could cause severe business damage, while the administration has defended the move in court filings.

Congressional interest in the company predates this week’s roundtable. The House Homeland Security Committee held a public hearing on December 17 on AI, quantum computing, and cloud security. The committee stated that the hearing focused on how emerging technologies can be used for defense and how foreign adversaries could exploit them, including limits on sharing sensitive technology abroad.

Anthropic has claimed that three Chinese AI companies used its Claude system to improve their own models, linking the claim to its push for tighter limits on sharing advanced technology.