Anthropic CEO Dario Amodei is scheduled to meet White House Chief of Staff Susie Wiles on Friday as the company seeks to resolve a deepening conflict with the Pentagon over access to its advanced artificial intelligence model, Mythos, according to a new report.
The meeting marks the most significant engagement yet between Anthropic and senior Trump administration officials since the dispute escalated earlier this year. According to Axios, the discussions reflect growing recognition within the administration that Mythos represents a major technological advancement, particularly in the field of cybersecurity.
Officials familiar with the matter told the outlet that the model's capabilities, including the ability to identify and potentially exploit weaknesses in digital systems, have made it both valuable and controversial. One source described denying the U.S. government access to such technology as strategically risky, citing competition with China.
The White House talks are unfolding against the backdrop of an ongoing legal and political battle between Anthropic and the Department of Defense. The Pentagon previously designated the company a "supply chain risk," effectively blocking it from defense-related contracts, as reported by Reuters. That designation followed Anthropic's refusal to allow its AI systems to be used without restrictions.
Anthropic has since challenged the move in court, intensifying tensions with military officials while continuing to engage with other parts of the federal government.
Despite the Pentagon's position, multiple civilian agencies are now seeking access to the Mythos model. According to Axios, departments including Energy and Treasury have expressed interest in using the system to assess vulnerabilities in critical infrastructure, such as power grids and financial networks.
Separate reporting indicates those efforts are gaining traction. The White House is considering granting federal agencies access to Mythos, even as the Pentagon maintains its blacklist, as reported by Bloomberg. The discussions include evaluating how the tool could be deployed safely within government systems.
An internal communication from the Office of Management and Budget confirmed that officials are reviewing agency requests to use the model, according to that report.
Anthropic has limited the release of Mythos to a small group of organizations, aiming to better understand its potential risks before broader deployment. Axios reports that the system's advanced cyber capabilities have raised concerns about how it could be misused if it fell into the wrong hands.
At the same time, those same capabilities are driving demand within government. Cybersecurity officials see the model as a way to simulate sophisticated attacks and strengthen defenses against real-world threats, particularly as concerns grow over foreign cyber operations targeting U.S. infrastructure.
The Pentagon, however, has maintained that restrictions imposed by Anthropic are too limiting. According to Reuters, defense officials have argued they need assurances that AI systems can be used for "all lawful purposes," a standard they say is incompatible with the company's current policies.
Anthropic has publicly stated that it will not allow its technology to be used for mass surveillance or the development of fully autonomous weapons, positions that have become a central sticking point in negotiations.
The standoff follows an earlier attempt to reach an agreement. In February, Defense Secretary Pete Hegseth gave Anthropic a deadline to accept the Pentagon's terms, a demand the company ultimately rejected, as reported by Axios. Since then, relations between the two sides have remained strained.
Even so, engagement has continued behind the scenes. Axios reports that Anthropic has brought on consultants with ties to the Trump administration, a move aimed at rebuilding communication channels and advancing discussions.
