BitGo chief executive Mike Belshe has issued an unusually literal challenge to Anthropic's AI model Claude: break into a Bitcoin wallet holding 100 BTC, and the coins are yours.
According to CCN.com, Belshe has framed the stunt as a direct test of whether AI can hack Bitcoin. Cryptonews.net reports that the BitGo CEO funded the wallet himself before issuing the dare, and bloomingbit describes the terms plainly: if Claude can move the 100 BTC, it can keep it.
Blockhead.co ties the challenge to a specific target — what it calls "Claude's hacking claims." In other words, this is less a general AI-safety exercise than a pointed response to the idea that frontier AI models are becoming capable offensive hackers. Startup Fortune and Crypto News both report the same core setup: real money, a public dare, and one of the industry's best-known custody executives putting his own holdings behind the argument.
The framing matters because BitGo is a crypto custodian — a company whose entire business is keeping other people's digital assets safe. A challenge like this doubles as a marketing claim about the strength of the cryptography Bitcoin wallets rely on, and about Belshe's confidence that no AI model, however advanced, can shortcut it.
It's worth noting what the available reporting does not establish: whether Anthropic has responded, whether anyone has actually attempted the challenge, or what rules govern it.
Why it matters: as AI companies talk up their models' growing cybersecurity abilities, a public, funded, winner-takes-the-coins test turns an abstract debate about AI hacking into a claim someone can actually try to falsify.