Ethereum Researcher: AI Could Break Crypto Wallets Soon

Ethereum Foundation researcher Justin Drake on Oct. 7 urged cryptocurrency holders and custodians to begin calm, deliberate planning for a potential AI-assisted cryptographic break of ECDSA — the elliptic-curve signature scheme that underpins many Ethereum and Bitcoin wallets. Drake framed this as a high-risk, low-probability scenario that could emerge faster than the classical quantum threat, suggesting a controlled migration to addresses whose public keys remain unexposed. Ethereum co-founder Vitalik Buterin supported taking AI-driven advances seriously but cautioned against a rushed mass migration that could create new security failures.

Why Drake is urging a move to fresh addresses

ECDSA (Elliptic Curve Digital Signature Algorithm) secures standard Ethereum accounts and many Bitcoin wallet types by enabling someone who holds a private key to sign transactions while the network verifies the signature against the corresponding public key. For addresses that have only ever received funds and never signed an outgoing transaction, the public key is not revealed onchain — only the address hash. Once that account signs a transaction, the public key can be recovered from the signature. If a future classical algorithm, potentially accelerated by AI, could derive the private key from the public key, any funds remaining at that address would be at risk.

What Drake recommends

Drake advised holders, beginning with large custodians and institutions, to plan a "bunker mode" migration. The recommended approach is gradual: move assets into addresses whose public keys remain hidden behind a hash (fresh addresses), prioritizing cold-storage and large balances first. Importantly, he suggested this migration could be done using addresses derived from the same seed phrase, so it does not necessarily require a new wallet format or an immediate cryptographic upgrade. Drake emphasized the need to avoid panic and poorly executed mass transfers that could create additional vulnerabilities.

How serious is the threat?

Drake characterized a worst-case scenario as an attacker being able to recover a private key from a public key in a timespan of roughly a week using large GPU clusters or similar classical hardware enhanced by AI-discovered algorithms. To date, no public attack has demonstrated such capability against ECDSA. The warning is therefore a forward-looking risk assessment that treats rapid advances in AI-assisted mathematics as a possible accelerant for cryptanalytic breakthroughs.

Bitcoin-specific nuances and Project Eleven

Not all Bitcoin addresses are equal when it comes to key exposure. Certain address types — including legacy P2PK (pay-to-public-key) and Taproot outputs — can expose public-key material even before an address spends, while reuse of addresses can expose keys for other formats. Project Eleven, which tracks onchain key exposure, reported more than 8.1 million BTC stored in addresses it classified as potentially quantum-vulnerable in its Sept. 14 update. That tracking is based on visible public-key exposure; being listed does not mean an address is already broken, just that it could be more exposed if a cryptanalytic shortcut appears.

Custodial cold storage under scrutiny

Drake named several large custodians — including Binance, Bitbank, Robinhood, Bitfinex and Tether — urging them to review cold-storage designs and address management. Institutions with large onchain balances or complex custody setups are a logical starting point for controlled migrations because they can coordinate secure processes and reduce the operational risks associated with any mass movement of funds.

Vitalik Buterin: take the risk seriously, but don’t rush

Ethereum co-founder Vitalik Buterin supported heightened attention to potential AI-driven cryptanalysis but warned users against immediate, hasty migrations. Buterin said he does not "recommend anyone scramble" to move funds now. His concern is broader than elliptic curves: he noted that lattice-based post-quantum schemes could also be affected by future AI-driven mathematical discoveries, recommending hash-based signature schemes where feasible and conservative parameter choices for lattice systems.

For individual holders, Buterin reiterated that keeping funds in addresses that have never signed can be a practical safeguard when done safely, but stressed that mistakes during rushed migrations — such as sending funds to the wrong address, exposing seed phrases, or creating reusable addresses — could cause more harm than the theoretical attack itself.

OpenAI publication and the prospect of accelerated math

Drake linked his timing to rapid advances in AI-assisted mathematics, calling out OpenAI’s Oct. 6 release of mathematical research produced by an internal frontier model. OpenAI published hundreds of manuscripts and proof material, arguing that the outputs vary in verification and formalization. The company did not claim to have broken ECDSA, RSA, or any cryptocurrency signature scheme.

Drake’s interpretation was more alarmist: he warned that "mathematical superintelligence" could enable new classical algorithms that significantly shorten the timeline for breaking elliptic-curve-based signatures, potentially arriving ahead of quantum hardware capable of running Shor’s algorithm at the necessary scale. Again, this represents a risk scenario rather than a demonstrated cryptanalytic break.

Ethereum’s post-quantum preparations and timeline

Ethereum has already been working on post-quantum resilience. The Ethereum Foundation created a dedicated Post-Quantum Security team in January 2026 and has been actively testing hash-based validator signatures, new proof systems, and cross-client interoperability for post-quantum solutions.

Key projects and standards

The network is developing leanXMSS, a hash-based signature scheme intended to replace quantum-vulnerable validator signatures, and leanVM, designed to efficiently aggregate larger post-quantum signatures. The target in public roadmaps is to have core post-quantum infrastructure ready around 2029, though Ethereum calls that a planning goal subject to change.

At the account and wallet level, EIP-8141 is intended to give accounts more flexibility over transaction authentication, with a roadmap placement targeting the Hegotá upgrade in 2027. That proposal could allow accounts to adopt quantum-resistant signature methods without forcing users to abandon existing account formats or move funds into entirely new addresses.

Outside of blockchain projects, the U.S. National Institute of Standards and Technology finalized FIPS 205 in 2024, standardizing SLH-DSA (based on SPHINCS+) as a hash-based digital signature option for post-quantum security. That provides an established, non-lattice-based alternative for systems seeking quantum-resistant signatures today.

Industry testing and institutional responses

Major custody providers and wallet vendors are already exploring post-quantum designs. BitGo tested post-quantum multi-party computation (MPC) signing with Silence Laboratories, and Coinbase is designing custody infrastructure that can adapt to whichever post-quantum signature standards gain prominence. These institutional experiments aim to minimize user friction while future-proofing key management and signature workflows.

Practical advice for holders and custodians

For retail holders and institutions alike, several practical steps align with both Drake’s recommendations and Buterin’s cautions:

• Assess exposure: Use onchain tools and trackers to determine whether addresses have already revealed public keys and which outputs could be more vulnerable.
• Plan controlled migrations: If a migration is appropriate, design it as a phased, auditable process starting with the largest balances and cold-storage keys to reduce operational risk.
• Prefer fresh address hygiene: Avoid address reuse, and prefer address types and wallet setups that hide public keys until a first spend reveals them.
• Coordinate custodially: Exchanges and institutional custodians should coordinate secure key rotation, tooling, and customer communication to reduce errors and abuse during any migration window.
• Watch for standards and upgrades: Track Ethereum’s EIPs, NIST standards like FIPS 205, and client releases that implement post-quantum validators or account-level upgrades.

Risk management, not panic

Drake’s warning is a call to risk management, not an alarm that ECDSA has already been broken. The scenario he describes would be severe, but it remains hypothetical until a reproducible cryptanalytic attack is published. The community response so far balances seriousness with caution: researchers and developers are accelerating post-quantum work while emphasizing careful operational practices to avoid self-inflicted losses during migrations.

Upcoming institutional forum

Drake is scheduled to present on post-quantum concerns to institutional participants on Nov. 12 in London during the Ethereum Institutional Forum. The agenda lists a 10 a.m. session titled "Post-Quantum Ethereum," followed by a technical roadmap discussion and live Q&A. That event will provide an important venue for institutional custodians, exchanges and developers to discuss practical steps and roadmap priorities for quantum-resistant signatures and wallet security.

Conclusion

The intersection of AI-driven mathematical research and classical cryptanalysis has shifted the conversation about when and how to harden blockchain systems against signature compromise. While no working attack has been published that breaks ECDSA or BLS signatures using classical hardware and AI, the potential for accelerated advances means developers, custodians and holders should take a pragmatic approach: assess exposure, plan controlled key rotations, follow emerging standards (like SLH-DSA/FIPS 205), and prioritize secure, coordinated action over panic. For now, the safest immediate practice for many users is straightforward: avoid address reuse, keep funds in addresses that have never signed where feasible, and follow custodial guidance for any large or complex holdings.