White House opposes Anthropic's expansion of Mythos usage to 120 companies, citing concerns over insufficient computing power

Odaily reported that the White House recently opposed Anthropic's proposal to expand the usage of its AI model Mythos to approximately 120 companies, primarily due to security and computing power concerns. Anthropic originally planned to add 70 new companies to the existing approximately 50 enterprise users of Mythos, but the White House questioned this, worrying that insufficient computing power could affect the government's own use of Mythos.

Mythos, launched in early April, is designed to detect and exploit critical software vulnerabilities and is currently limited to testing by enterprises managing critical infrastructure, with no plans for public release. The White House fears that if usage is expanded to more commercial users, the government will face computing power bottlenecks when using the model. This is particularly concerning given that while Anthropic has signed computing power procurement agreements with Amazon, Google, and Broadcom, new capacity has not yet come online.

On the political front, relations between the White House and Anthropic have not improved. The Trump administration has publicly criticized Anthropic for hiring multiple former officials from the Biden administration and has expressed dissatisfaction with its ties to liberal organizations. One example highlights the trust issues between the two sides: Anthropic former researcher Collin Burns was originally assigned to a government AI model evaluation role, but senior White House officials replaced the candidate upon learning of the arrangement, citing the need to avoid conflicts of interest by preventing AI company personnel from directly engaging in matters involving other AI companies.

Additionally, Anthropic disclosed an unauthorized access incident involving the Mythos model last week, further intensifying external regulatory scrutiny of the company.