On the eve of the listing, there are signs of easing tensions between Anthropic and the Trump administration after months of disputes.

According to informed sources, as AI unicorn Anthropic prepares to go public, signs of easing tensions between Trump administration officials and the company have been emerging in certain departments of the US government. Earlier this year, the relationship between Anthropic and the US government broke down after Anthropic refused to allow the US military to use its AI models for domestic surveillance and fully automated weapon systems. The government retaliated by placing the company on a national security blacklist set to take effect later this year. In March, the US Department of Defense labeled the company as a "supply chain risk," marking the first time a US company has received this designation typically reserved for enterprises with ties to hostile nations. This label prohibits tens of thousands of contractors from using Anthropic's AI products while working for the US military. Since the dispute erupted, relations have improved somewhat after Anthropic CEO Dario Amodei visited the White House in mid-April and discussed potential collaboration. However, Anthropic is still challenging the "supply chain risk" designation in court. Franklin Turner, a lawyer specializing in government contracts, stated that the Department of Defense is vigorously defending its position against Anthropic's lawsuit. He mentioned that the broader damage to the company's business is unlikely to be alleviated until the dispute with the Department of Defense is fully resolved. Turner said, "Any time the government sends a signal to 'divest itself' from a company, it's a big problem for that company." One of the biggest symbolic but crucial signs of the warming relationship is that, according to two informed sources, the White House had invited Amodei to the signing ceremony for an executive order on artificial intelligence originally scheduled for May 21. However, the event was canceled later due to Trump's dissatisfaction with certain provisions in the order. Trump signed the executive order on Tuesday, and Anthropic posted a statement on social platform X expressing anticipation to "work together" with the White House to implement the order. Opening Channels with the White House According to an informed source, Anthropic has been in discussions with National Cyber Director Shawn Kane-Cros about its advanced AI system "Mythos," exploring ways to protect critical infrastructure from AI-initiated cyberattacks. The company had previously warned that Mythos could significantly enhance the potency of such attacks. Critical infrastructure refers to important sectors of the US economy that should be safeguarded from attacks, such as banks, emergency services, and hospitals. As Anthropic's relationship with the Trump administration gradually warms, the company is also preparing to go public, with a potential valuation of $1 trillion. CEOs of large companies have been working to strengthen their relationships with the White House to avoid upsetting Trump. Harrison Rolfs, a senior research analyst at PitchBook focusing on AI companies, stated that closer ties can help boost investor confidence in the short term. Rolfs referred to this dispute as a "short-term bruise." According to a US official, Anthropic employees met with Treasury Secretary Scott Bennett this spring to discuss Mythos and potential actions the president could take regarding AI. These discussions reportedly helped Trump's team in drafting the executive order issued on June 2, which requires leading developers to submit their advanced models for cybersecurity testing. Tensions with the Pentagon have not completely dissipated. As of Thursday (when both sides submitted their defenses to the court), they were still disputing the designation of "supply chain risk." Meanwhile, US Army spokesperson Sean Mindon stated that Anthropic employees did not participate in an AI simulation exercise led by the Army on April 27 targeting cyberattacks. According to an Army statement, network executives from leading AI developers like Google and OpenAI were involved in the exercise.