Anthropic has been blacklisted by the US government, downstream companies such as Figma (FIG.US) face business risks.

Anthropic The dispute between PBC and the US government over whether the company's AI model will be banned by federal agencies has posed a financial threat to other companies. Design software manufacturer Figma (FIG.US) stated that if Anthropic continues to be considered a "supply chain risk," it may harm its ability to sell products to government agencies. In a regulatory filing on Thursday, Figma disclosed that Claude, an AI feature built by Anthropic, is relied upon for products sold to federal agencies. The filing stated that if the US government blacklists Anthropic and Figma cannot find suitable alternatives, "sales to government agencies and highly regulated organizations may be impacted." The US Department of Defense's contradictions and fears about Anthropic In February of this year, US President Donald Trump stated that due to disputes with Anthropic over the military's use of its model, the US government would blacklist the company as a "supply chain risk." Subsequently, Anthropic sued the Department of Defense, claiming that its ban was because the company opposed the way the technology might be deployed. It is worth noting that in addition to the commercial disputes on the surface, the US political circle's deep fear of Anthropic's internally developed supermodel with the codename "Claude Mythos" is also a fundamental factor driving the Department of Defense to take extreme measures. According to multiple Wall Street sources and intelligence sources, the "Mythos" model has shown shocking and chilling high talents in the field of cybersecurity. It can automatically discover "zero-day vulnerabilities" in highly complex large-scale distributed software systems without any human intervention and generate extremely accurate exploit tools. This top-tier hacker-level automated destructive power has deeply alarmed decision-makers, including US Vice President Pence and White House cybersecurity advisors. Washington is generally concerned that once this technology with the potential for autonomous weaponization loses absolute control, if malicious hostile forces exploit it or the model itself deviates logically, it could cause catastrophic disasters to public hospitals, bank clearing systems, and the national power grid in the US. Although the Department of Defense is currently deploying the model to patch security gaps in its own systems due to network defense pressure, this has only strengthened its determination to completely remove this "uncontrollable" enterprise from the core supply chain through administrative means. Anthropic being blacklisted, downstream companies like Figma "step on landmines" collectively With the model provider being classified as a supply chain security risk, Figma is likely to face catastrophic blows such as federal order losses in the short term, or even the cancellation of existing government customers collectively withdrawing and unsubscribing. What concerns the senior management even more is that this panic is spreading to highly government-compliant industries such as finance, healthcare, and utilities. Due to the time and difficult to estimate research and development costs required to find, test, and seamlessly replace an equivalent large-scale language model, Figma's core product's commercial realization and revenue growth expectations have been seriously overshadowed in the absence of timely technological decoupling. It is understood that Figma's disclosure echoed similar risk alerts issued by other companies earlier. Network security provider Tenable Holdings Inc. also informed investors earlier this month that rapid replacement of Anthropic's products may lead to "significant engineering costs, service interruptions, and loss of key product features." Global shipping platform Freightos Ltd. stated last month that its use of Anthropic's model faced "significant regulatory and political risks" due to a legal dispute with the Department of Defense. Figma is a company that provides software for website and application design, and it was authorized to sell products to federal agencies in early 2025. According to the submitted documents, the company expects government sales to increase over time, but government sales currently do not account for a high proportion of its revenue. Figma's nightmare reveals the harshest survival rule of the AI era While Anthropic is being pushed to the forefront due to adhering to security bottom lines, the fierce commercial competition in the AI race has not stopped. Its biggest rival, OpenAI, quickly seized this policy dividend. Within hours of Anthropic being excluded from the procurement list, OpenAI reached an agreement with the Department of Defense, and by the end of February 2026, signed a strategic acquisition agreement to introduce advanced AI models into military secret networks, rapidly seizing the market advantage. Although the initial agreement contained broad terms allowing for "all legal purposes" and sparked internal controversy, OpenAI subsequently urgently amended the contract in early March, clearly defining three security red lines of "prohibition for domestic large-scale surveillance, prohibition of independently guiding autonomous weapons, and prohibition of automated high-risk decisions." In addition, the market vacuum was not ultimately monopolized by OpenAI. In May, the Department of Defense included eight tech giants including OpenAI, Alphabet Inc. Class C, NVIDIA Corporation, and xAI as suppliers for its secret network. Meanwhile, the alliance relationship between Figma and Anthropic has recently shown subtle and dangerous cracks. It is worth noting that the Chief Product Officer of Anthropic, who had long held a key position on the Figma board, officially stepped down in April, and shortly afterwards, Anthropic made a high-profile launch of the "Claude Design" native product that directly targeted the design industry. This move has made the capital market acutely aware that SaaS software companies not only have to bear the political and compliance risks brought by the underlying model providers but also have to be on guard against these industry giants who possess core underlying technologies evolving into their most destructive direct competitors at any time.