Buterin: AI Won’t Break Bitcoin, 90% of Net Worth in Crypto

Ethereum co-founder Vitalik Buterin rejected a claim that artificial intelligence could cause Bitcoin to lose more than half its value within two years. He placed a low probability on AI breaking Bitcoin’s hashing algorithms or its proof-of-work consensus and said most AI-driven security problems should be fixable. Buterin noted his holdings — about 90% of his net worth — already reflect a bet on that outcome.

The warning that prompted the exchange came from an investor who argued there is a significant chance AI could undermine investor confidence by exposing practical weaknesses in crypto infrastructure. That argument does not require AI to break core cryptography; it relies on a pattern of attacks that reveal exploitable flaws and shake market trust.

Buterin framed the issue as a question of whether developers and defenders can patch and harden systems faster than attackers can find and exploit weaknesses. He pointed to the distinction between attacks on surrounding infrastructure — wallets, bridges, exchanges and client software — and attacks that would break the blockchain’s core consensus rules.

Industry evidence shows attackers using automation and AI techniques to probe systems. In August, a Bitcoin swap service suspended operations after months of automated, AI-assisted probing produced several contained exploits. The service said its non-custodial design protected customer funds, but the company absorbed losses and concluded it could not continue operating safely.

Security specialists highlight that many components that hold or move crypto assets present more entry points than the blockchain protocol itself. Deddy David, chief executive of blockchain security firm Cyvers, warned the financial exposure from AI-assisted attacks could reach “hundreds of millions or billions of dollars” if models identify vulnerabilities at scale.

Technology companies are developing defensive tools in response. One company limited public access to a highly capable model and redirected its use to a program that scans for and repairs vulnerabilities with the help of major cloud providers and financial firms. That effort includes commitments of usage credits for defensive work.

More than 100 organizations, including leading AI developers and technology firms, joined an open letter urging governments and companies to build advanced defensive AI and provide security testing for critical infrastructure such as hospitals and utilities. Banks and payment firms also called for stepped-up protections.

The central open question for crypto remains whether attackers can gain a decisive advantage before defensive tools, auditing practices and network infrastructure catch up. Crypto systems hold transferable assets and operate on public networks, which can magnify the impact of successful attacks and shorten the window for remedial action.