Key Highlights
- Federal authorities mandated Anthropic to disable Fable 5 and Mythos 5, its most sophisticated AI systems, based on national security justifications.
- Officials cited concerns about a potential jailbreak technique, though Anthropic characterizes it as limited in scope and ineffective as a widespread vulnerability.
- While Anthropic followed the government’s instructions, the company has openly challenged the directive, arguing the supporting evidence is inadequate.
- The company’s pre-IPO trading instruments on Hyperliquid declined 3.7%, settling near $1,627 after the announcement.
- Anthropic has committed to publishing comprehensive technical information regarding the alleged security exploit within one day and maintains the directive results from miscommunication.
Federal authorities commanded Anthropic to deactivate its most capable artificial intelligence systems on Friday, sending shockwaves through pre-IPO trading markets and establishing a contentious standoff between the AI company and government regulators.
Details of the Federal Directive
The government’s mandate reached Anthropic at 5:21 p.m. Eastern Time on Friday evening. Exercised through national security provisions, the instruction demanded the organization terminate all foreign national access to Fable 5 and Mythos 5, regardless of their physical location.
This restriction extended to foreign nationals employed within Anthropic’s own workforce.
For full compliance, [[LINK_START_0]]Anthropic[[LINK_END_0]] chose to completely disable both systems for its entire user base. The company’s remaining offerings, including Opus 4.8, continued operating without interruption.
Both Fable 5 and Mythos 5 had only recently become available to users. These systems were developed using Mythos Preview as their foundation—a versatile model the company credited with identifying thousands of security flaws in essential software infrastructure.
Federal officials informed Anthropic they had discovered a technique capable of circumventing Fable 5’s protective measures. Such jailbreak methods enable users to override built-in safety restrictions within AI models.
After examining the technique, Anthropic challenged the government’s assessment. According to the company, this approach represents a targeted exploit rather than a universal bypass mechanism, and depends on a handful of previously identified, relatively insignificant vulnerabilities.
The company further noted that other commercially available models, including OpenAI’s GPT-5.5, can identify identical vulnerabilities through standard operations without requiring any circumvention techniques.
“Up to this point, federal authorities have only provided verbal testimony regarding a potentially limited, non-universal exploit method, which fundamentally involves instructing the model to analyze particular source code and identify any security weaknesses,” Anthropic stated.
The organization emphasized that cybersecurity professionals routinely perform this identical function.
Market Reaction to the Shutdown
Financial markets reacted swiftly to the development. Anthropic’s perpetual futures contract on Hyperliquid, serving as an indicator of projected equity valuation, dropped 3.7% by Saturday to approximately $1,627.
This trading instrument had exceeded $1,800 following Fable 5’s initial release. Current open interest stands at roughly $8.6 million—modest when compared to contracts like SpaceX, yet significant for an organization that hasn’t initiated public offering procedures.
Anthropic maintains its position that the government’s action stems from a fundamental miscommunication and is actively pursuing a swift resolution to reinstate system access.
The organization intends to publish complete technical documentation concerning the purported security exploit within the next 24 hours.
Anthropic cautioned that implementing this benchmark universally “would effectively freeze all future model releases across all leading AI developers.”
The organization, which established its market position through an emphasis on responsible AI development, now finds itself in the unusual position of openly contesting a federal security mandate based on insufficient supporting documentation.
The eventual outcome—whether authorities withdraw, modify, or expand the directive—will significantly influence Anthropic’s trajectory toward becoming a publicly traded entity.
Get 3 Free Stock Ebooks
Discover top-performing stocks in AI, Crypto, and Technology with expert analysis.
- Top 10 AI Stocks - Leading AI companies
- Top 10 Crypto Stocks - Blockchain leaders
- Top 10 Tech Stocks - Tech giants
