Silicon Valley vs. The Oval Office: White House Halts Anthropic’s ‘Mythos’ Expansion

WASHINGTON / SAN FRANCISCO — The White House has formally opposed plans by AI startup Anthropic to expand access to its groundbreaking “Mythos” model to an additional 70 companies, according to a report by the Wall Street Journal.

The intervention highlights a deep-seated mistrust between President Donald Trump’s administration and the California-based developer behind the Claude chatbot, potentially stalling a major leap in global cybersecurity infrastructure.

The Power of ‘Mythos’

Anthropic claims its “Mythos” model is a generational talent in the world of code. Unlike previous models, it can reportedly identify undiscovered security loopholes that have remained hidden for decades—vulnerabilities that have eluded both the best human hackers and current automated tools.

• The “Glasswing” Project: Anthropic has been extremely cautious, sharing a limited version only with “tech titans” like Apple, Microsoft, and Nvidia to harden their defenses.

A Marriage of Inconvenience

The relationship between the White House and Anthropic has been rocky at best.

• The Military Dispute: Tensions flared when Anthropic refused to grant the U.S. military “unconditional use” of its software.
• The “Supply Chain Risk”: In February 2026, Pentagon chief Pete Hegseth officially designated the company a “national security supply chain risk,” leading President Trump to order the government to stop using their tech immediately.

Why the White House Said ‘No’

The administration’s opposition to the 120-company expansion rests on two main pillars:

• Hackers & Security: Authorities fear that wider access increases the “surface area” for an attack, worrying that the model itself could be exploited by foreign adversaries.
• The Compute Crunch: Interestingly, the government is worried about computing power. They fear that if Anthropic shares its “compute” with 120 different companies, there won’t be enough processing power left to support the government’s own priority AI projects.

The Courtroom Battle

Anthropic isn’t backing down. The company is currently fighting the government’s “supply chain risk” designation in court, arguing that its tools are essential for national defense, not a threat to it.