The Zhitong Finance App learned that AI company Anthropic PBC (ANT.US) is about to conduct a landmark initial public offering (IPO). On the eve of listing, the company issued a warning to investors in its prospectus: since the company had previously been outspoken on AI security issues, it may be the target of government retaliation. This highlights that after a year of friction with the Trump administration, the company's future business is facing unusual government-related risks.
Prospectus warns: may be “selectively targeted” by the government
Anthropic said in a prospectus submitted on Monday that the company's position on AI regulation issues and statements made by executives may put it under tremendous pressure from the government. The prospectus stated, “We have experienced and may continue to experience higher levels of attention, and may also continue to be subject to stricter scrutiny by specific government agencies, or face the risk of being disproportionately and selectively targeted, even if other similarly situated companies are not treated the same way.”
These statements highlight the tense relationship between the company and the Trump administration. The tension began earlier this year over a dispute over whether the Pentagon should put in place protective measures when using the company's technology. Recently, Anthropic also clashed with officials over whether its cutting-edge model, Mythos, is safe enough and can be publicly released. Furthermore, the company's CEO Dario Amodei supports coordination to slow down AI development so that security teams can keep up with technology. This position also continues to cause dissatisfaction with US President Trump.
Investors face pricing challenges: regulatory uncertainty is hard to measure
Investors must now decide how to weigh the regulatory risks this company faces. Anthropic's technology is both close to everyday consumers, can be packaged as everyday products, and is powerful enough to be viewed as a cyber weapon and potential biological warfare catalyst, and therefore requires greater scrutiny. For Anthropic, repeated disputes with the government have further exacerbated the uncertainty that has engulfed the entire AI industry.
Sarah Kreps, director of the Science Policy Research Institute at Cornell University, said, “In the field of AI, the boundaries of government power are still being clarified, sometimes through direct confrontation with individual companies. It's hard for investors to factor this uncertainty into pricing.”
In the prospectus, Anthropic revealed various potential government-related setbacks and mentioned federal legislative proposals, including mandatory shutdown models and even nationalization of the AI business. According to reports, earlier this year, some officials, including Vice President Vance, proposed using the Cold War “Defense Production Law” to force companies to disclose information and cooperate with national security agencies.
Anthropic also listed other risks that government intervention could affect its AI development efforts, including trade controls that could cut off its ties to foreign customers, and the legal consequences of misbehaving in its models. The company said it may not be able to “fully understand or explain” the behavior of the system it is developing and will continue to push for regulation that could hinder its own business.
The company said, “The regulations or other measures we promote, if passed, will apply to ourselves, and may increase our compliance burden, limit our flexibility in product development or deployment, affect how and when we develop and release new models, or otherwise slow our growth.”
Anthropic also warned that the technology it is developing could eventually pose a threat to humans, in line with many Silicon Valley AI researchers' concerns about the “end of AI,” which Trump and his allies dismissed as a “scam.” Anthropic said, “Advanced AI systems can pose a disastrous or existential risk if proper assurance, supervision, and security measures are not taken during development, deployment, and operation.”
Anthropic said recent initiatives to link AI models to scientific research may trigger restrictions aimed at preventing the manufacture of lethal biological weapons. The company also added that its model has been misused in cyber attacks, which could expose the company to legal liability and reputational damage.
Behind the dispute: the government is both a regulator and a customer
As highlighted by the Pentagon dispute, the government is not only a regulator, but also an important customer of the company's products, which further constrains it to the will of the government. The company said direct contracts from government agencies accounted for about 1% of its revenue during the various periods covered by the document (which can be traced back to 2023).
Former US Deputy Secretary of Industry and Security for Export Control William Reinsch said, “Anthropic is a bit special because in the Trump administration, many actions are made on a whim, and Anthropic isn't treated the same way as other companies.”
The Pentagon dispute was sparked by Anthropic's insistence on imposing additional safeguards on its technology, which eventually led US Secretary of Defense Hegseth to announce in March this year that the company posed a supply chain risk and banned it from participating in military work. Although Anthropic has filed a lawsuit to reverse this finding, an appellate court in Washington upheld the Pentagon's plan to remove Claude from its system and require defense contractors to do the same. Other civilian government agencies are not bound by Hegseth's decision.
Anthropic said in the document: “We are uncertain whether the outcome of these lawsuits will affect our ability to sell products to government agencies and our ability to sell products to customers who provide products and services to the US government, thereby adversely affecting our business, financial situation, and operating results.”
Anthropic has also received additional scrutiny from the Trump administration for its cutting-edge model Mythos. The company warned in April of this year that the model is extremely good at finding computer vulnerabilities, so it can only be shared with a few trusted partners. This breakthrough prompted US Treasury Secretary Bezent to inform Wall Street about potential hacking risks, and eventually prompted Trump to sign an executive order on June 2 stipulating a voluntary evaluation system for the new cutting-edge model.
Due to differences with the government over securing the new system, the US Department of Commerce briefly imposed export controls on Mythos 5 and Fable 5, a consumer-oriented related product, forcing the company to temporarily close all access rights. Two weeks later, export restrictions were lifted after Anthropic addressed a jailbreak concern that could give Fable users access to Mythos abilities. However, Anthropic warned in its prospectus that future export controls will not necessarily be lifted on a similar schedule, and that future growth will “require” international expansion.
Since then, the US government has increased its focus on AI safety. Trump summoned more than 20 Silicon Valley leaders to the White House last month to urge AI developers, including Amodei, to sign voluntary agreements, calling on companies to revise board structures, add dedicated security teams, and conduct independent audits of models. A few days later, Trump administration officials issued legal threats through the Federal Trade Commission, implying that companies deemed to be in violation of the agreement could face consequences. The move raised concerns about government intervention.
US officials, particularly those in the Trump administration, are increasingly tempted to take more measures to restrain AI companies, which in itself is unprecedented. This situation is even more unusual for a company that is at the center of most relevant debates and is going public at a time when the US is trying to regulate an industry that is still in its infancy.
Matt Pearl, former White House National Security Council director and currently leading strategic technology programs at the Center for Strategic and International Studies, said, “We don't know what kind of organization Anthropic and other leading AI companies will eventually become. I've never seen a situation where the future of a large company will be and what kind of relationship it will have with the government is so uncertain.”