Anthropic PBC confirmed Tuesday that a strict US government order requiring the immediate shutdown of its Mythos AI model will not disrupt service for the handful of venture capital firms and corporate partners who were already using it.
The artificial intelligence startup assured its earliest backers that the federal mandate, which cited severe national security risks, is exclusively intended for the broader population.
The Commerce Department issued an emergency directive last week demanding a total halt to the Mythos system, citing unmitigated capabilities that could threaten critical infrastructure. Anthropic moved swiftly to comply with the federal order, immediately revoking access for academic researchers, safety auditors, and waitlisted developers. However, the company quietly kept the model's application programming interface fully operational for select financial institutions and early corporate testers.
We completely agree with the government's assessment that this technology is far too dangerous for widespread deployment, which is why we are restricting its use solely to our most lucrative corporate clients.
Several early users, including private equity firms and algorithmic trading desks, reported no interruption to their Mythos preview access. The firms have been utilizing the pre-release model to automate high-frequency trades and generate synthetic market research, applications that Anthropic executives have described as vital to the broader innovation ecosystem. The Commerce Department has yet to issue enforcement penalties against the company for maintaining the active server connections.
Federal regulators have historically granted enforcement leeway to technology companies that successfully restrict their most catastrophic products to enterprise pricing tiers.