Anthropic's Pre-IPO Shares Fall as US Government Shuts Down Its Most Powerful AI Model
The U.S. government ordered Anthropic to shut down access to its Fable 5 and Mythos 5 AI models due to national security concerns over a potential jailbreak, causing a 3.7% drop in Anthropic's pre-IPO shares on Hyperliquid.
Intelligence analysis by Gemini 2.5 Flash
Anthropic's shares tumbled after the US government mandated a shutdown of its most powerful AI models, Fable 5 and Mythos 5, citing national security concerns. The company disputes the directive, arguing the vulnerability is narrow and commonly found in other leading AI models, potentially hindering future AI development.
Imagine a super-smart computer brain called Anthropic made two amazing helper programs, Fable 5 and Mythos 5. The grown-ups in charge in the US told Anthropic to turn them off because they found a tiny secret backdoor, like a hidden button, that someone could use to make the helpers do things they shouldn't. Anthropic thinks it's a small problem, and other smart computer brains have similar hidden buttons, but now people are worried about how much Anthropic's company is worth.
Analysis
The U.S. government issued a directive to Anthropic on Friday, June 13, 2026, at 5:21 p.m. ET, demanding the immediate shutdown of its two most advanced AI models, Fable 5 and Mythos 5. The order, based on national security authorities, initially required suspending access for foreign nationals but was later expanded to a full shutdown for all customers.
The government cited a method to 'jailbreak' Fable 5, allowing bypasses. Anthropic, however, contested this, stating that the identified technique was a narrow exploit targeting minor, previously known vulnerabilities. The company highlighted that similar vulnerabilities could be found in other publicly available models, including OpenAI's GPT-5.5, without needing a bypass.
Anthropic expressed strong opposition to the government's standard, arguing that applying such a strict criterion across the industry would 'essentially halt all new model deployments for all frontier model providers.' This dispute is particularly notable given Anthropic's brand focus on safety-first AI development. The company plans to release more technical details about the specific jailbreak within 24 hours.
The market reaction was immediate and negative. Anthropic's perpetual contract on Hyperliquid, which tracks expectations for the company's equity price, fell by approximately 3.7% to about $1,627 on Saturday. This drop from its post-launch highs above $1,800 has sparked new concerns regarding Anthropic's pre-IPO valuation and its future public listing ambitions. The outcome hinges on whether the government's order is reversed, narrowed, or extended following Anthropic's technical rebuttal.
Key points
- The U.S. government ordered Anthropic to shut down its Fable 5 and Mythos 5 AI models due to national security concerns.
- Anthropic's pre-IPO shares on Hyperliquid fell 3.7% following the government's directive.
- Anthropic disputes the order, claiming the jailbreak is narrow and similar vulnerabilities exist in other AI models like OpenAI's GPT-5.5.
- The company argues that the government's standard could halt future AI model deployments across the industry.
- Anthropic plans to release technical details about the jailbreak within 24 hours.
If Anthropic successfully demonstrates the narrowness of the vulnerability and the government reconsiders its directive, the company could recover its market position and proceed with a strong IPO. This situation could also lead to clearer, more collaborative regulatory frameworks for AI safety, benefiting the entire industry.
The government's order could be upheld or expanded, severely impacting Anthropic's ability to innovate and deploy new models. This could permanently damage its IPO prospects and valuation, while also creating a chilling effect on AI development across the industry due to overly stringent national security concerns.
