Ethereum co-founder Vitalik Buterin has endorsed a warning from Ethereum researcher Justin Drake that fast-moving artificial intelligence could weaken the cryptography protecting blockchains, possibly before quantum computers become a practical threat. The concern is focused on the elliptic curve digital signature algorithm (ECDSA), which secures cryptocurrency wallets.
What prompted the warning
Drake spoke up after OpenAI published hundreds of new mathematical findings in areas such as algebra, theoretical computer science and mathematical logic. According to the report, OpenAI also used a team of 10,000 autonomous AI agents to work on the Navier-Stokes equation, a famously difficult unsolved problem, over 88 hours. Drake said recent AI results have challenged long-held mathematical assumptions, which could expose weaknesses in elliptic curve cryptography.
- Drake said elliptic curves have a rich mathematical structure that could make them more exposed to AI-driven attacks.
- He contrasted them with hash functions, which are built to have little algebraic structure and are seen as safer for now.
- Buterin extended the worry to lattice-based cryptography, which is designed to resist quantum attacks. He suggested the next two years could bring significant reductions in its concrete security.
Suggested precautions
Both men said holders could consider moving funds to fresh addresses whose public key has never been revealed. Drake said large and sophisticated holders should go first, and that any remaining assets should move to a new address after a transaction is signed. Buterin said that if it is not difficult, keeping funds in unused addresses is a sensible precaution. He also cautioned that migrations should be gradual and carefully managed, because rushed moves could create new risks.
Haseeb Qureshi, managing partner at Dragonfly, echoed the call for caution. He said the risk comes not from quantum computing but from ordinary mathematics overturning cryptographic assumptions. The reported consensus among researchers is that immediate action is not required, but the community should take AI-related cryptographic risk seriously and prepare for possible changes in security practices.
Why it matters
Blockchains rely on math problems that are believed to be too hard to solve in practice. A wallet's private key is protected by that difficulty, and ECDSA is the signature scheme behind it. Most public discussion of this risk has centered on quantum computers. This warning points to a different path: better AI tools finding shortcuts in existing mathematics.
For someone learning about crypto, it shows that security rests on assumptions that researchers keep testing. It also explains why an address whose public key has not been exposed is considered more protected, and why experts favor slow, careful upgrades over panicked moves. The warning is a precaution from researchers, not a report of a successful attack.




