US Government Bans Anthropic’s Fable 5 and Mythos 5 Over Security Concerns

The Trump administration forced Anthropic to pull two new models citing national security, but researchers and the company itself are pushing back on the rationale.

The United States government has ordered Anthropic to withdraw its two newest artificial intelligence models, Fable 5 and Mythos 5, from availability, citing national security concerns. [1]

The action followed reports that Amazon researchers allegedly discovered a method to bypass the safety guardrails built into Fable 5. [1]

Anthropic has responded by noting that the same jailbreak techniques — methods used to circumvent a model’s built-in restrictions — exist in other AI models as well, not just its own. [1]

Cybersecurity researchers have also pushed back, signing an open letter characterizing the government’s move as dangerous. [1]

The ban has immediate practical consequences for developers building products and services on Anthropic’s platform, as access to the two newly released models has been suspended. [1]

The episode is being described as the latest development in what observers have characterized as a fraught relationship between Anthropic and the Trump administration. [1]

Amazon’s chief executive reportedly raised concerns about the Anthropic models before the government crackdown took effect. [1]

Some analysts have suggested the ban could paradoxically benefit Anthropic’s public profile, a dynamic that has drawn attention ahead of the company’s anticipated initial public offering (IPO). [1]


Sources

  1. TechCrunch — Is the US government’s Anthropic ban accidentally helping the brand?

This article was drafted with AI from the cited sources and checked against them before publication. Spot an error? Let us know.

Sam Altman told staff to expect a public offering 'within the next year,' signaling a likely slip from earlier timelines despite a prospectus already on file.

OpenAI chief executive Sam Altman told employees via Slack that the company expects an initial public offering (IPO) — a first sale of shares to the public — “within the next year,” suggesting the listing could slide into 2027 rather than happening imminently. [1]

Altman characterized the already-filed prospectus as a way to preserve “optionality if we want to go sooner,” rather than a signal of an imminent listing. [1]

The CEO offered two competing forces that could influence timing: progress on self-improving AI could push the IPO back because “technology and the world may change in surprising ways, and there might be good reasons to be a private company during that time,” while large capital requirements for compute infrastructure could accelerate it. [1]

The delay comes as rival AI lab Anthropic appears set to go public in the coming weeks, and OpenAI is still burning cash — factors that could weigh on OpenAI’s valuation if it listed now. [1]

Alongside the IPO update, Altman announced an upcoming secondary stock sale for employees priced at $687.69 per share. [1]

On the product side, OpenAI is preparing a new model internally codenamed 5.6, which research lead Jakub Pachocki describes as a significant step up from GPT-5.5, with a possible June release. [1]


Sources

  1. The Decoder — OpenAI's IPO slips as Altman tells staff to expect a public offering "within the next year"

This article was drafted with AI from the cited sources and checked against them before publication. Spot an error? Let us know.

OpenAI is in talks to lease a massive federally-sited data center in Ohio, with Nvidia reportedly set to guarantee the lease and financing.

OpenAI is negotiating to lease a planned 10-gigawatt data center in southern Ohio that could be financially backed by Nvidia, according to two people with direct knowledge of the talks cited by The Information. [1]

The site is located on federal land in Pike County, Ohio — the former home of a uranium enrichment facility — and is being developed by SB Energy, a company majority-owned by SoftBank, which is also an investor in OpenAI. [1]

Under the proposed terms, OpenAI would sign a 20-year lease, which would represent its largest infrastructure commitment to date. [1] Total costs at full buildout are expected to reach at least $500 billion. [1]

Nvidia’s role would go beyond supplying chips: the company would act as guarantor for both the lease and the project’s financing, backing payments with its own balance sheet. [1] That level of financial guarantee would be unprecedented for the chipmaker at this scale. [1]

The first phase of the project, at 800 megawatts of capacity, is expected to be operational by 2028. [1] Negotiations are described as ongoing, and plans could still change. [1]

The project draws comparisons to the Stargate initiative — announced at the White House in January 2025 and involving Oracle and SoftBank — which ultimately made little progress. [1]

Separately, OpenAI on Monday confidentially filed paperwork for an initial public offering (IPO). [1]


Sources

  1. The Decoder — OpenAI wants its biggest data center yet, and Nvidia would back the bill

This article was drafted with AI from the cited sources and checked against them before publication. Spot an error? Let us know.