A warning from Ethereum researcher Justin Drake has triggered a debate over whether artificial intelligence could accelerate a breakthrough against the cryptography securing crypto wallets and blockchains. Drake says the industry should prepare for a possible failure of ECDSA, the elliptic-curve signature system widely used across crypto, potentially within months rather than years in a worst-case scenario.
According to Cointelegraph, Drake’s proposed “bunker mode” would involve moving assets to fresh addresses whose public keys have not been exposed. The aim would be to reduce the information available to an attacker attempting to derive private keys from public keys. Critics, however, say the approach is incomplete, difficult to implement and potentially dangerous if it encourages hurried transfers.
Why the warning has caused concern
The scenario under discussion is broader than the long-standing risk posed by quantum computers. Drake believes advanced AI could help researchers discover mathematical attacks using conventional computers—methods that human cryptographers have not yet found.
The concern focuses on elliptic-curve cryptography, which underpins ECDSA and many blockchain signatures. Drake argues that elliptic curves contain more mathematical structure than hash functions and could therefore be more exposed to a major mathematical breakthrough. His warning described a worst-case possibility in which ECDSA fails before a so-called “Q-day”, the point at which quantum computing is powerful enough to threaten current cryptographic systems.
The assessment has not been accepted across the industry. Matthew Green, a cryptography professor at Johns Hopkins University, said public-key cryptography could potentially be lost. By contrast, Yehuda Lindell, Coinbase’s head of cryptography, said there was, to his knowledge, no evidence that the mathematical assumptions behind elliptic-curve cryptography had weakened.
Lucien Bourdon, a Bitcoin analyst at hardware-wallet maker Trezor, also said there were no credible published advances against the mathematics behind ECDSA, whether produced with AI or through other means. He noted that recent AI-generated mathematical papers had not produced an ECDSA attack. Bourdon also questioned Drake’s argument about elliptic curves, saying the researcher had not explained how that additional mathematical structure would lead to a break.
Why a fresh address may not solve the problem
Moving coins to a new address can appear to offer protection because an unused address may not yet have exposed its public key. But that protection depends on what information has already been shared with wallets or third parties.
Zach Herbert, chief executive and co-founder of hardware-wallet company FoundationHQ, said an extended public key, or xpub, can allow a wallet to generate multiple public keys for the same account, including addresses that have never been used. If an application or service already has the xpub, it may be able to identify the supposedly hidden address. In that situation, moving funds within the same account would not necessarily conceal the relevant information.
Eli Ben-Sasson, co-founder and chief executive of Ethereum scaling protocol StarkWare, said simply moving coins would not be sufficient. Users would also need new types of wallets if they intended to continue using their assets after a cryptographic failure. The consequences could extend beyond wallet signatures: Ethereum’s infrastructure signatures, consensus mechanisms and KZG polynomial commitments used for data blobs would also require attention. Ben-Sasson added that gas prices could rise sharply if substantial network functionality stopped working.
Taylor Monahan, founder and chief executive of Ethereum wallet MetaMask, made a broader point: if elliptic-curve cryptography failed and financially motivated attackers could compromise exposed keys, the disruption would affect much more than crypto prices. In such a systemic event, placing coins in a “bunker” would not by itself preserve the wider ecosystem.
Different networks have different defenses
Bitcoin address types do not all expose public keys in the same way. Some addresses use a hash of the public key and reveal the full key only when funds are spent. If the relevant hash functions remain secure, this can provide an additional barrier. Common SegWit addresses beginning with “bc1q” keep public keys hidden until spending, while wallets such as Trezor generate new receiving addresses to reduce reuse.
However, spending Bitcoin exposes the public key in the mempool, creating a possible window for an attack. Bourdon said there was no immediate action required and warned that rushed transfers could create mistakes and give scammers an opportunity to exploit fear.
Other networks are pursuing different approaches. NEAR Protocol allows users to add a quantum-safe FIPS-204 key to an existing account in one transaction, while its Meteor wallet supports quantum-safe signing. Ethereum is working on a post-quantum roadmap involving hash-based signatures and zero-knowledge-proof aggregation. A possible emergency hard fork has also been outlined, including measures to disable conventional wallet transactions and allow recovery through zero-knowledge proofs of a seed phrase. That proposal has no set timetable.
Starknet’s account-abstraction design allows users to change signature schemes without moving funds, although other parts of its architecture and its reliance on Ethereum would still need review. Solana exposes public keys by default, meaning a simple move to a new address would not address the issue. Its foundation has described a possible upgrade using hash-based proofs to let users migrate accounts to safer signatures, but that upgrade has not been deployed.
Current market picture
Moneyiar market data at 21:23-21:24 on October 15, 2026, shows mixed performance among the listed assets:
- JUST (JST): $0.1324452848, down 6.28% over 24 hours and 3.89% over seven days.
- AKEDO (AKE): $0.0332164248, up 25.95% over 24 hours and 27.22% over seven days.
- RedStone (RED): $0.1582799004, up 0.73% over 24 hours and 9.15% over seven days.
- Propy (PRO): $0.3719933611, up 2.25% over 24 hours and 9.71% over seven days.
The experts quoted in the report do not describe an active ECDSA attack. Max Shannon of Bitwise Europe said the theoretical risk was worth treating prudently, while also indicating that Bitwise was not currently focused on it. The central disagreement is therefore about preparation, not evidence that cryptography has already failed. For now, critics argue that an unplanned migration could create more immediate operational and security risks than the unproven threat itself.
