Ethereum co-founder Vitalik Buterin said on October 7 that there is a "good chance" AI-driven advances in mathematics could compromise lattice-based cryptography within the next two years, and he advised builders to prefer hash-based signature schemes wherever feasible.

The warning targets a family of post-quantum cryptography that the industry had largely treated as durable. Buterin named two systems at particular risk: ML-DSA, a digital signature scheme designed to resist quantum attacks, and Fully Homomorphic Encryption (FHE), a technique that allows computation on encrypted data without decryption. Both rely on lattice mathematics.

What did Buterin say about ECDSA?

Beyond lattice schemes, Buterin flagged elevated risk to ECDSA signatures, the mechanism crypto wallets use to authorize transactions. He noted the danger grows as public keys become exposed on-chain, since an address that has sent a transaction reveals the public key and widens the attack surface.

Ethereum researcher Justin Drake made a parallel argument the same day. Drake said recent breakthroughs in AI-driven mathematics could surface vulnerabilities in ECDSA much sooner than previously anticipated, and he called for faster preparation.

Buterin framed his position as a probability, not a certainty. The two-year window is a forecast, and no concrete AI-assisted result against lattice schemes or ECDSA currently exists. Any such breakthrough would shift his warning into an operational emergency.

What should holders and builders do?

Buterin urged users not to make panic-driven moves with their assets, but he offered two concrete recommendations:

• Prefer hash-based signatures over lattice schemes wherever feasible.
• Use fresh addresses that have never executed a transaction, because addresses that have never sent funds have not yet revealed their public keys.

For builders, the guidance carries design consequences. Projects relying on lattice-based schemes for signatures or encryption may need to revisit those choices, particularly where hash-based alternatives are already workable.

No immediate market impact followed the statements from either Buterin or Drake. The operational takeaway for individual holders is address hygiene: limiting on-chain exposure of public keys reduces the information available to a future attacker.

How does this align with Ethereum's roadmap?

Buterin's comments track a direction Ethereum had already taken. The network's Lean roadmap has shifted toward hash-based signature schemes such as WOTS and SPHINCS+, aiming to cut dependence on cryptographic methods vulnerable to quantum attacks, AI-driven attacks, or both.

The exposure also extends beyond blockchains. The same cryptographic assumptions underpin secure websites, messaging applications, and VPNs, meaning an AI-driven break in lattice schemes or ECDSA would have consequences across mainstream internet infrastructure.

For now, the two-year window functions as a planning horizon rather than a deadline. The next concrete signal to watch is any demonstrated AI-assisted cryptanalytic result against ML-DSA, FHE, or ECDSA, which would convert Buterin's forecast into an urgent migration question for wallets, protocols, and the broader internet stack.