Ethereum Foundation researcher Justin Drake is urging the crypto industry to begin preparing for what he calls “bunker mode,” warning that advances in AI-assisted mathematics could eventually undermine the elliptic-curve cryptography protecting major blockchain wallets. Drake is calling for precautionary planning rather than claiming that ECDSA has already been broken, with his proposed defense centered on gradually moving high-value holdings toward addresses whose public keys have not been exposed.
The warning follows a sharp advance in machine-assisted mathematics. On October 6, OpenAI published hundreds of mathematical results produced by an internal frontier model, including computer-checkable Lean formalizations for many proofs. Drake cited developments like these as evidence that AI could discover mathematical shortcuts faster than previously expected. His “months, not years” scenario remains a speculative worst case, not a demonstrated cryptographic capability.
Unexposed Public Keys Provide a Temporary Defense
Drake’s proposed bunker strategy relies on an existing property of some blockchain addresses: the public key can remain hidden behind a hash until a transaction reveals it. If elliptic-curve signatures were broken, exposed public keys could theoretically become starting points for private-key recovery. Moving funds methodically to fresh, unused addresses could reduce exposure without requiring an immediate new cryptographic standard.
The protection is not universal. Ethereum externally owned accounts reveal enough signature information after transacting to recover their public key, while Bitcoin address types vary; Taproot outputs, for example, expose public keys differently and use Schnorr rather than ECDSA. Both systems ultimately rely on elliptic-curve discrete-log assumptions. The issue has already pushed Bitcoin custodians toward post-quantum migration planning and prompted practical quantum-resistance experiments on Bitcoin. Address hygiene can buy time, but it is not a permanent substitute for cryptographic migration.
Vitalik Buterin broadly agreed that AI-driven mathematical progress deserves attention while warning against an emergency transfer campaign. He said keeping funds in unused addresses is sensible when easy, but noted that migration mistakes themselves can destroy assets. Operational risk therefore becomes part of the security calculation: reducing cryptographic exposure is useful only if the migration process does not introduce a larger failure mode.
Ethereum Pushes Toward Hash-Based Security
Ethereum was already pursuing post-quantum defenses before Drake’s latest warning. The Foundation formed a dedicated Post-Quantum Security team in January and is developing hash-based signatures, signature aggregation and protocol changes intended to reduce dependence on elliptic curves. The broader Ethereum post-quantum roadmap currently targets core infrastructure around 2029, although those milestones remain provisional. Drake’s AI concern argues for accelerating an existing migration rather than inventing a new security program from scratch.
Hash-based cryptography is attractive because it removes the additional structured mathematical assumptions behind elliptic curves and lattice schemes. Drake has increasingly favored that approach, while Ethereum research has also reconsidered which hash functions should anchor the transition, including its recent shift away from Poseidon for parts of the Layer 1 roadmap. The objective is not merely quantum resistance, but cryptographic foundations considered harder for future AI systems to shortcut mathematically.
For now, there is no evidence that AI can recover production Bitcoin or Ethereum private keys from exposed public keys. The significance of Drake’s warning is therefore the compression of the planning horizon, not confirmation of an active attack. Custodians, exchanges and protocol developers now face a strategic question: how quickly to improve key hygiene and cryptographic agility without turning a hypothetical future threat into an immediate operational risk through rushed migrations.
