Anthropic chief executive Dario Amodei met White House chief of staff Susie Wiles and Treasury Secretary Scott Bessent on Friday in what both sides described as a productive step towards resolving the AI company's bitter dispute with the Trump administration.
The meeting, first reported by Axios, represents the highest-level engagement between Anthropic and the White House since Defence Secretary Pete Hegseth designated the company a supply chain risk in early March, a label previously reserved for firms linked to foreign adversaries.
The White House said the talks were "productive and constructive" and that opportunities for collaboration had been discussed alongside approaches to balancing innovation and safety.
Anthropic said the meeting explored how the company and the US government could work together on national security and economic priorities.
The encounter was driven by the administration's recognition that Anthropic's new model, Mythos, unveiled on 7 April, presents capabilities too significant to ignore, even as the Pentagon continues to pursue its legal campaign against the company.
Mythos has demonstrated the ability to surpass human cybersecurity experts in finding and exploiting software vulnerabilities, prompting Anthropic to restrict its release to a select group of organisations under a controlled initiative called Project Glasswing.
Partners including Amazon, Apple, Google, Microsoft and JPMorgan Chase are using the model to identify weaknesses in their own systems before malicious actors gain access to similar capabilities.
A source close to the negotiations told Axios it would be "grossly irresponsible" for the US government to deny itself the advantages the model offers, describing inaction as a gift to China.
Bessent joined the meeting because he wanted to ensure all parties were aligned, according to a source familiar with his thinking, reflecting Treasury's concern that Mythos-class tools could be used to breach the US financial system if defenders are not given access first.
The Office of Management and Budget told cabinet department officials earlier in the week that it was putting protections in place to allow federal agencies to begin using Mythos, Bloomberg reported, while the New York Times' DealBook newsletter said the Treasury and State Departments had separately requested briefings and access.
The dispute between Anthropic and the Pentagon erupted in late February after Hegseth demanded unrestricted access to the company's Claude models for all lawful purposes, including potential use in autonomous weapons and domestic surveillance.
Amodei refused, and the Pentagon designated Anthropic a supply chain risk, the first time an American company had received the label.
Anthropic filed two federal lawsuits alleging illegal retaliation, and a judge in California blocked the wider government ban, though a Washington appeals court subsequently allowed the Pentagon to maintain its own designation while litigation continues.
Co-founder Jack Clark said this week that Anthropic was discussing Mythos with the administration despite the ongoing legal fight, and warned the world should prepare for more powerful systems.
Related reading
- OpenAI closes $122bn funding round at $852bn valuation as IPO pressure builds
- OpenAI strengthens Sora video safety and consent controls
- Anthropic launches Claude Design
"There will be other systems just like this in a few months from other companies, and in a year to a year-and-a-half later, there will be open-weight models from China that have these capabilities," he told the Semafor World Economy conference.
The meeting signals a potential pathway for Anthropic to resume working with civilian agencies while the Pentagon dispute plays out separately in the courts.
The recap
- Amodei to meet White House chief of staff Susie Wiles.
- Mythos used in Project Glasswing for defensive cybersecurity access.
- White House says new technology requires security evaluation period.
