US Lifts Anthropic AI Export Ban as Bitcoin Holds Near $59K

Crypto News

The United States has lifted its export controls on Anthropic's Claude Fable 5 and Mythos 5, clearing both frontier AI models on June 30 and paving the way for global access to resume from July 1. The Commerce Department's decision ends nearly three weeks of tense negotiations between the AI developer and the White House. Anthropic confirmed the reversal directly, stating it had received notice that the restrictions were removed and would begin restoring access the following day. The move dissolves a regulatory standoff that had knocked two of the company's flagship products offline worldwide since mid-June, disrupting enterprise and developer workflows across dozens of markets.

The original directive landed on June 12, just three days after both models went live on June 9. It cited national security concerns reportedly tied to potential model jailbreaks — attempts to bypass an AI system's built-in safeguards. The order required Anthropic to suspend access for any foreign national, inside or outside the country, including its own foreign-born staff. It marked the first such command the administration had issued against a frontier AI model, and it arrived at a delicate moment as the company prepared for a public listing. Company leadership travelled to Washington to press its case with White House officials.

Enforcing the ban proved unworkable in practice. Segmenting users by nationality in real time was effectively impossible, so Anthropic pulled both models entirely offline rather than risk non-compliance. The suspension spanned every product surface — the Claude.ai consumer platform, the developer API, Amazon Web Services' Bedrock, and other partner deployments. For nearly three weeks, customers who had wired the models into production systems were left without access. The episode underscored how abruptly regulatory intervention can sever a widely used software service, and how little operational flexibility a developer retains once a national-security order is issued.

Claude Fable 5 ranks as Anthropic's most capable widely available model, built on the firm's Mythos-class architecture but shipping with enhanced safeguards for general use. It targets demanding reasoning tasks, long-horizon agentic work, software engineering and advanced vision. Mythos 5 shares the same underlying architecture but carries relaxed safeguards for cybersecurity work, and access was restricted to trusted partners through a program called Project Glasswing. The company had judged Mythos 5 too proficient at finding and exploiting security vulnerabilities for broad release. Such frontier systems increasingly underpin crypto tooling, from an AI trading bot to an AI crypto wallet.

The timing was costly. Anthropic filed a confidential S-1 draft with securities regulators on June 1, taking its first formal step toward an initial public offering that markets expect could arrive this year. The three-week blackout cost the company momentum on its flagship products just as rivals advanced — OpenAI rolled out its GPT-5.6 family, while Alphabet's Google pushed its Gemini 3.5 Pro release into July. Restoring access lets Anthropic re-enter the frontier-model race. On the same day the controls lifted, the company also unveiled Claude Sonnet 5 and Claude Science, a triple product launch signalling it intends to reclaim lost ground quickly.

Friction with the administration is far from settled. In March, US Defense Secretary Pete Hegseth formally designated Anthropic a supply-chain risk, a ruling the company is challenging in a lawsuit that remains before the courts. Commerce Secretary Howard Lutnick said his department had spent two weeks analysing and approving Fable 5 to meet government standards, though neither side disclosed what technical safeguards were added. The administration has also published a new AI executive order letting leading developers voluntarily submit models for review up to 30 days before public release. Chinese rivals such as Alibaba, meanwhile, gained ground during the blackout window.

COINOTAG's reading: this saga sits at the intersection of AI policy and risk-asset sentiment, and its resolution lands in a jittery market. Our aggregate data shows the Fear & Greed Index at 11 out of 100 — deep in extreme-fear territory — with Bitcoin dominance at 69.7% and total crypto market capitalisation near $1.70 trillion. With Bitcoin trading around $59,000, capital is concentrating in the majors while the broader altcoin complex bleeds. Clearer, faster regulatory pathways for frontier AI reduce one tail risk hanging over technology valuations, and the AI-crypto narrative — from autonomous agents to on-chain compute — remains among the few themes still attracting flows in an otherwise defensive tape.