Anthropic's Mythos Made Everyone Hit the Panic Button—Even the Fed Is Taking Notes

Anthropic has officially become the AI equivalent of that friend who shows up to parties uninvited but somehow everyone's still glad they came—or at least, everyone's watching them like a hawk wondering if they'll accidentally burn the house down. The vibes are mixed, to say the least.

In what might be the most awkward power lunch of the quarter, U.S. Treasury Secretary Scott Bessent and Federal Reserve Chair Jerome Powell sat down with senior Wall Street executives at the Treasury Department. The menu probably featured light refreshments and heavy concerns about Anthropic's newest AI model, Mythos, and the cybersecurity nightmare it allegedly brings wrapped in a very shiny bow.

According to Bloomberg, regulators are out here telling major banks to fortify their digital castle walls because Mythos apparently has a talent for spotting and exploiting vulnerabilities across major operating systems and web browsers when prompted. It's the kind of feature that looks absolutely mesmerizing in product demos but makes security professionals reach for their therapy appointments.

Not one to wait for the inevitable regulatory slap on the wrist, Anthropic went ahead and restricted access to the model while simultaneously launching Project Glasswing—a defense initiative that reads like a who's-who of corporate heavy hitters including Amazon, Apple, JPMorgan Chase, Google, NVIDIA, and Linux Foundation. The whole point? Use Mythos to find and patch vulnerabilities in systems where the stakes are higher than a degen's life savings on a random Tuesday.

The company is flexing hard, claiming Mythos has already uncovered thousands of high-severity flaws across widely used software environments. In actual commitment terms, that's $100 million in usage credits and $4 million in direct funding for open-source security efforts. More than 40 organizations maintaining core infrastructure have been handed the keys for what they call defensive applications—because apparently, the best defense against a digital wrecking ball is having it on your payroll.

Financially, Anthropic is absolutely not struggling. Annual recurring revenue went from $9 billion at the start of 2026 to $30 billion—a number so wild it would make even the most bullish crypto degens do a double-take. That's the kind of growth that makes boardrooms collectively lose their minds.

On the legal front, because why not add more drama to this whole situation, a federal court in San Francisco blocked the Pentagon's designation of Anthropic as a supply chain risk. Judge Rita Lin basically told the Pentagon that their designation lacked statutory support and called the whole thing arbitrary—which is legal speak for "y'all didn't do your homework."

Despite market whispers suggesting Anthropic might be coming for firms like Palantir Technologies, analyst Dan Ives was quick to pour cold water on those concerns, pointing to strong revenue growth in Palantir's U.S. commercial and government segments. So the Palantir bulls can rest easy