Ethereum co-founder Vitalik Buterin has urged holders not to rush digital-wallet migrations, even as he treats faster AI-driven mathematical research as a genuine risk to the cryptography that secures wallets and chains.
His comments, posted on X on October 7, 2026, came in reply to Ethereum researcher Justin Drake, who had called for a controlled “bunker mode” after recent AI systems produced a burst of new mathematical results.
Drake’s concern is that machine-assisted math could weaken the elliptic-curve digital signature algorithm (ECDSA) that underpins Bitcoin and Ethereum wallets sooner than many security models assume.
In that framing, a public key revealed when an address first spends funds becomes a longer-term exposure.
Drake suggested a gradual shift of assets toward addresses that have never broadcast a transaction, starting with larger and more sophisticated holders, and moving residual balances after each spend.
He also warned that a hurried migration could do more harm than good.
Buterin agreed that the industry should take AI-accelerated cryptanalysis seriously and reduce reliance on schemes that might prove fragile, including some post-quantum designs based on lattices.
I don't recommend anyone scramble to move their funds to new wallets today. But we should take the risks to cryptography from AI-accelerated math seriously, and minimize our exposure to not just quantum-vulnerable cryptography, but also potentially AI-vulnerable cryptography.…
— vitalik.eth (@VitalikButerin) October 7, 2026
He noted that a common assumption—elliptic curves eventually broken, hashes and lattices remaining safe—may not hold if the next couple of years of AI-assisted math erode the concrete security of lattice constructions.
That view, he said, is one reason Ethereum’s lean roadmap has leaned toward hash-based designs where those are practical.
He also favored more conservative parameters for any lattice system, off-chain delivery of encrypted notes in privacy protocols, and off-chain multisig confirmations so that signer keys are not published on-chain.
If ECDSA were compromised earlier than expected, an off-chain multisig could degrade to the coordinator’s key rather than becoming spendable by anyone.
The operational advice was deliberately restrained.
Buterin said he does not recommend that anyone scramble to new wallets today.
Keeping funds in unused addresses is reasonable when it is easy, he wrote, but migrations themselves are hazardous.
He said he has personally lost more money to botched moves than to every hack combined.
Dragonfly’s Haseeb Qureshi separately described Drake’s note as a sober warning that ordinary mathematical progress, not only quantum computers, could overturn unproven hardness assumptions.
The security discussion landed during a sharp market drawdown.
On October 7, ether slipped below $2,600, with reported prints near $2,555 to $2,616, while bitcoin traded around $84,000 after a multi-day retreat.
CoinGlass-based tallies put crypto-wide forced liquidations in a range of roughly $550 million to $665 million over 24 hours, the large majority of them long positions.
Ether accounted for on the order of $175 million to $201 million of that total, and one Binance ETH-linked position of about $26.6 million ranked among the largest single closures.
Elevated open interest and previously positive funding had left leveraged bulls exposed when prices reversed.
Reports also cited ether ETF outflows and a jump in oil prices tied to tanker attacks near the Strait of Hormuz as part of the broader risk-off backdrop.
However, it shoudl be clear now that the two stories are only loosely connected.
Buterin’s note is a medium-term cryptographic planning argument, not a claim that wallets are broken now, and the sell-off is a leverage unwind. Together they underline a familiar tension: theoretical key risk can tempt holders into hasty operational changes at the same moment that volatile markets are already closing out crowded long bets.