Anthropic regains model access after U.S. lifts export controls

Traders Union
All News
Financial News
Anthropic regains model access

Highlights

The U.S. Department of Commerce lifted export controls on Anthropic's Fable 5, allowing its public relaunch after the model's June 12 suspension.
The export ban and its reversal intensify disputes between the White House, tech sector, and foreign governments over AI regulation and cross-border access.
OpenAI's GPT-5.6 faces similar U.S. restrictions, with initial access limited to two dozen government-approved partners and a broader release likely next week.

Ban reversal clears path for model relaunch

As first reported by Financial Times, the U.S. Department of Commerce tells Anthropic it has lifted export controls on the company’s leading models, according to people with knowledge of the matter. The decision, communicated on Tuesday evening, allows the AI group to re-release Fable 5 to the general public.

The government already allowed Anthropic last week to re-release Mythos 5, a model with fewer safeguards aimed at enterprise customers, to about 100 pre-vetted partners. Commerce secretary Howard Lutnick imposed the controls on June 12 after officials found a jailbreak that bypassed the model’s safety guardrails, prompting Anthropic to withdraw the system for users worldwide.

Anthropic declines to comment. The company becomes the first AI group to face a White House ban after concerns that its technology could be used to exploit cyber security vulnerabilities in critical industries.

AI policy tensions extend to allies and rivals

The episode heightens tensions between the White House and the technology sector, which criticizes what it sees as an ad hoc approach to regulating frontier AI. It also triggers backlash from foreign governments, which accuse the U.S. of penalizing allies by extending the ban beyond its borders.

The export control follows the Trump administration’s rollout of a voluntary framework for monitoring advanced models, while also pledging not to impose a formal regulatory regime on leading AI companies. At the same time, several senior U.S. officials warn against broad public release before the systems’ ability to exploit cyber security weaknesses is properly assessed.

OpenAI faces similar limits on GPT-5.6, with the San Francisco-based company asked to restrict initial access to roughly two dozen government-approved partners. A broader release could come as early as next week, according to a person familiar with the matter, while the U.S. and Europe are also discussing a trusted partner scheme that would give close allies preferred access to the latest AI tools.

Our earlier article on the ICTS Supply Chain Security Act explained how U.S. senators proposed to codify and expand the Commerce Department’s authority to review and restrict foreign-linked technology used in critical communications and AI-related infrastructure. We noted that the bill would add enforcement tools and create a clearer framework for prohibiting certain technology transactions tied to countries of concern, while also including safeguards for free speech and open-source software.

This material may contain third-party opinions, none of the data and information on this webpage constitutes investment advice according to our Disclaimer. While we adhere to strict Editorial Integrity, this post may contain references to products from our partners.