Vitalik Buterin is confident AI won't break crypto, and he's putting his money where his mouth is — roughly 90% of his net worth, in fact. The Ethereum co-founder's stance runs against a growing chorus of warnings that AI could undermine the security guarantees that underpin Bitcoin and other proof-of-work networks.
The 90% bet
Buterin's position is not a casual opinion. His existing crypto holdings effectively bet 90% of his net worth on the outcome. He argues that Bitcoin can handle security failures that don't require social consensus, and places a low probability on AI breaking hashing or proof-of-work. In his view, the core cryptographic foundations are robust enough to withstand AI-driven attacks.
The counterpoint
Not everyone shares that confidence. Liron Shapiro predicts a 50% probability that Bitcoin will fall more than 50% within two years, citing AI's potential to undermine security guarantees. The argument goes that AI could find weaknesses in the network's assumptions faster than humans can patch them. Buterin's response is that Bitcoin's design already accounts for certain types of failures — and that the ones AI might trigger aren't the ones that require social consensus to fix. It's a disagreement about where the real vulnerability lies.
Real-world attacks
The debate isn't purely theoretical. Bitcoin swap provider Boltz suspended service in August after months of automated, AI-assisted probing led to exploits. The company's non-custodial architecture protected customer funds, but Boltz absorbed the losses itself. The timing isn't great for skeptics: the outage came just last month. Deddy David, CEO of Cyvers, said AI-powered crypto attacks could eventually reach hundreds of millions or billions of dollars in financial exposure. That's a far cry from the theoretical risk Buterin is betting against.
The defensive push
Meanwhile, the industry is mobilizing. Anthropic restricted public access to its Claude Mythos model after it demonstrated the ability to autonomously discover and weaponize software vulnerabilities. The company launched Project Glasswing with partners including AWS, Google, Microsoft, and JPMorgan, committing up to $100 million in usage credits. More than 100 organizations, including Google, Microsoft, Anthropic, and OpenAI, are involved in the defensive push against AI-driven attacks. The scale of the response suggests the threat is taken seriously at the highest levels of tech and finance.
Buterin's bet is a public one. The industry's answer will come in the form of whether Bitcoin's security holds as AI tools get cheaper and more capable. For now, the Ethereum co-founder is holding his ground — and his crypto.
