A crypto executive has issued an unusual public challenge to Anthropic: come and take it.
According to a report from BeInCrypto, surfaced via Google News, the CEO publicly dared the AI company to hack a Bitcoin wallet holding roughly $6.3 million. The dare frames the stash as a live target — an open invitation to prove that a frontier AI model can, or cannot, crack the cryptography protecting a real wallet with real money in it.
Beyond the challenge itself and the size of the prize, details remain thin. The available reporting does not establish whether Anthropic has responded, what rules or timeline the wager would follow, or how anyone would verify an attempt. Readers should treat the specifics as unconfirmed until fuller accounts appear.
The stunt lands on a live nerve, though. Bitcoin wallets are secured by cryptography whose whole premise is that no attacker — human or machine — has enough computing power to guess a private key. AI companies, meanwhile, have spent the past few years publishing research on how capable their models are at finding software vulnerabilities, a capability that cuts both ways for defenders and attackers. A multimillion-dollar public bounty turns that abstract debate into a scoreboard.
It matters because the answer, whichever way it goes, is a rare public data point on whether today's AI systems are anywhere near threatening the math that holds up crypto — and much of the internet.