Officials object to opening the model to about 70 more companies after Anthropic earlier let roughly 50 organisations test Mythos under an early-access programme.
The White House cited national security and computing-capacity concerns, while Anthropic said it is in productive talks with the government and disputed claims the model would drain resources.
The dispute follows Anthropic's disclosure of a possible unauthorised Mythos access and wider tensions with the Pentagon over military uses of its AI systems.
Why is the government blacklisting Anthropic while seeking exclusive access to its most controversial AI model?
Is Anthropic's 'dangerous' AI a revolutionary weapon or an overhyped tool with its own glaring security flaws?
White House Blocks Anthropic’s Mythos AI Expansion Amid Security Breach and Pentagon Supply-Chain Risk
Overview
In early 2026, the Pentagon labeled Anthropic a national security supply-chain risk after failed talks demanding unrestricted access to its powerful Mythos AI, leading to a government ban and legal battles. Despite this, key agencies like the NSA continue using Mythos for cybersecurity, creating resource strains and ethical concerns over the government's exclusive access. A recent breach of Mythos through a third-party vendor intensified fears, prompting the White House to block Anthropic's civilian expansion plans while crafting guidance to bypass Pentagon restrictions for government use. This contradictory approach fuels regulatory confusion, stifles broader innovation, and deepens geopolitical tensions in global AI governance.