Technology & Science

U.S. export-control order shutters Anthropic’s Fable/Mythos AI worldwide, then White House quickly back-pedals

After a June 12 Commerce Department order barred any foreign access to Anthropic’s new Fable 5 and Mythos 5 models, the firm pulled the systems offline globally; within a week President Trump told Axios he no longer sees the company as a security threat, signalling a likely restoration.

By Underlines Team

Focusing Facts

  1. Directive issued 12 June 2026 under Export Administration Regulations forced Anthropic to disable Fable 5 and Mythos 5 for all users after a 90-minute compliance window.
  2. On 19 June 2026 Trump said Anthropic had “behaved very responsibly” and that he is “not sure” emergency Defense Production Act powers are needed, softening the earlier stance.
  3. Anthropic, valued at $965 billion, had confidentially filed its IPO paperwork on 1 June 2026, meaning the shutdown hit less than two weeks before a planned US$60 billion listing.

Context

Washington last tried to muzzle software at scale during the 1990s crypto-export fight (Bernstein v. DoJ, 1995-99), but unlike compiled ciphers that ship on disks, frontier AI now lives in the cloud—so export controls instantly cascade to every overseas keyboard. The episode exposes two structural currents: (1) the state is expanding Cold-War style trade rules from physical chips to intangible model weights, and (2) mega-labs financing multibillion-dollar compute bills via IPOs are becoming systemically important—inviting antitrust and security scrutiny reminiscent of Standard Oil’s 1911 breakup or the 1946 Atomic Energy Act’s nationalization of fission know-how. That the White House reversed tone in a week, apparently prodded by an Amazon complaint and G7 diplomacy, shows both how improvisational present governance is and how little sovereignty corporations really enjoy. A century from now this may mark an early skirmish in the ‘splinternet of intelligence,’ where nations demand kill-switches on cloud AI just as they once demanded backdoors in encryption, accelerating the geographic fracturing—and relocation—of model development.

Perspectives

Libertarian commentary outlets

Counter Punch, Crypto BriefingPortray the Commerce Department’s export-control order as an unconstitutional attack on free speech and proof that AI firms should flee U.S. jurisdiction. Their long-standing hostility toward government regulation leads them to dismiss the national-security rationale and exaggerate the ease of operating offshore, even though neither article shows concrete evidence the models pose zero risk.

Financial-investor media

Goodreturns, MintFocus on Anthropic’s extraordinary revenue growth and looming trillion-dollar IPO, treating the government clash mainly as a material risk factor for shareholders rather than a policy problem. The investor lens incentivises upbeat coverage that hypes valuations and market potential while skimming over ethical or security concerns raised in the same articles.

Government-aligned / national-security-minded outlets

The Hindu, Deseret NewsEcho the administration’s narrative that export controls were a responsible step to protect U.S. security, with Trump now satisfied that Anthropic is cooperating. By foregrounding official statements and applauding Washington’s newfound ‘big stick,’ these pieces risk normalising heavy-handed intervention and downplaying the competitiveness or free-speech costs highlighted elsewhere.

Like what you're reading?

Create a free account to read 5 articles every week. No credit card required.

Share

Related Stories