The growing conflict between Anthropic and the US government is starting to create problems far beyond the AI company itself. What began as a dispute over military use of AI models is now turning into a business risk for software firms that rely on Anthropic’s technology, especially companies working with federal agencies.
Figma recently warned investors that the ongoing dispute could hurt its government business if Anthropic continues to be labeled a “supply chain risk” by US authorities. The issue matters because Figma uses Anthropic’s Claude models to power AI features offered to federal customers.
The situation traces back to February, when the Trump administration moved to blacklist Anthropic from certain government work after disagreements over how the company’s AI systems could be used by the military. Anthropic later sued the Defense Department, arguing that the restrictions were unfair and politically motivated.
What makes this story important is the ripple effect. Companies that integrated Claude into enterprise products are now realizing they may need backup AI providers. That is not a quick switch. Replacing a foundational AI model often requires retraining workflows, rebuilding integrations, retesting security controls, and rewriting compliance documentation.
Figma is not alone. Cybersecurity firm Tenable and logistics platform Freightos have also disclosed concerns tied to Anthropic dependency.
The bigger issue here is how deeply AI vendors are now embedded inside enterprise software stacks. Many SaaS companies market their own AI tools, but the underlying models often come from a small group of providers like Anthropic, OpenAI, or Google. If one provider faces regulatory trouble, customers downstream can also feel the impact.
For investors, this adds a new layer of IPO and enterprise risk analysis. It is no longer enough to ask whether a company has AI features. Regulators, procurement teams, and shareholders are increasingly asking whose AI models those features depend on.
