The BRICS TimesThe Trump administration has raised serious security objections to Anthropic’s plan to open up its most advanced AI model, Mythos capable of hacking major software systems to dozens of new companies and organisations.
White House Draws a Hard Line on Mythos Expansion
The Trump administration has pushed back hard against Anthropic’s move to widen access to its most powerful artificial intelligence model, Mythos. The AI company had proposed giving around 70 additional companies and organisations access to the tool. That would have brought the total number of entities using Mythos to roughly 120. White House officials made clear in private conversations they strongly disagreed with the plan.
The concerns, first flagged by a leading financial newspaper, come from unnamed sources close to the discussions. Anthropic has not responded publicly, and the White House did not immediately clarify its position.
Why Mythos Has Everyone on Edge
Mythos is no ordinary AI tool it is, by Anthropic’s own admission, a world-class cyber weapon. The company says the model has already uncovered “thousands of high-severity vulnerabilities, including some in every major operating system and web browser.” Many of those flaws had gone undetected for decades. That extraordinary capability is exactly what makes the White House nervous about putting it in more hands.
Officials fear that broader access could enable large-scale cyber attacks should the technology fall into the wrong hands. Beyond the threat of misuse, there is another concern: computing capacity. White House officials worry Anthropic may not have enough processing power to serve 120 organisations without compromising the government’s own access to the model.
A Fraught Relationship Between Anthropic and Washington
This standoff is the latest chapter in a tense, months-long feud between Anthropic and the federal government especially the Pentagon. The trouble started when Anthropic refused to sign an agreement allowing the military to use its Claude AI for “all lawful purposes.” The company drew two firm red lines: it would not allow its technology to support mass domestic surveillance of Americans, or to power fully autonomous weapons systems.
The Pentagon wanted no such restrictions. It called Anthropic an unreliable partner and issued a “supply chain risk” designation an unprecedented move that effectively blacklisted the company from many federal systems. Defense Secretary Pete Hegseth formalised the ban in February 2026, directing agencies to stop using Anthropic’s AI.
The courts stepped in, offering Anthropic partial relief. A federal judge called the administration’s measures “Orwellian” and said they could “cripple” the company. A higher court, however, declined to block the Pentagon’s blacklisting while the legal battle plays out.
A Possible Path Back But Tensions Remain
Despite the friction, there are signs the two sides may be looking for an exit ramp. The White House is said to be drafting guidance that could allow civilian agencies to work around Anthropic’s supply-chain risk designation and adopt Mythos. One source described the effort as a practical way to “save face and bring em back in.”
Earlier this month, White House Chief of Staff Susie Wiles and Treasury Secretary Scott Bessent met Anthropic CEO Dario Amodei in what both sides described as a productive introductory meeting. President Trump himself said on CNBC’s Squawk Box that Anthropic was “shaping up” and, when asked about a potential Pentagon deal, added: “It’s possible. We want the smartest people.”
Still, even if the Pentagon lifts its blacklisting which some insiders believe will happen eventually the core disagreement over usage restrictions has not gone away. OpenAI and Google have both signed deals allowing the Pentagon to use their models under the “all lawful purposes” standard. Anthropic has not.
Mythos Already Inside Some Government Systems
Ironically, despite the feud, Mythos is already operating inside parts of the U.S. government. The National Security Agency has integrated the model into certain analytical systems a detail that highlights just how complicated and contradictory this standoff has become.
The broader picture is this: multiple federal agencies are eager to access Mythos, even as the Pentagon fights Anthropic in court. And Anthropic itself is said to be in talks with investors to raise funds at a staggering $900 billion valuation more than double its current worth partly to build enough computing infrastructure to meet that surging demand.
