Ethereum Researcher’s ‘Bunker Mode’ Call Sparks Debate Over Whether AI Math Threatens Crypto Keys
Ethereum researcher Justin Drake warns that AI‑driven math could soon threaten ECDSA, urging a cautious shift to unused addresses, but the community remains divided on the urgency and practicality of such a move.

The “Bunker Mode” Debate: AI Math vs. Crypto Keys
In early October, Ethereum Foundation researcher Justin Drake issued a warning that AI‑driven mathematics could soon threaten the elliptic‑curve signature scheme (ECDSA) that underpins most Bitcoin and Ethereum accounts. He urged holders to move funds to unused addresses—a strategy he dubbed bunker mode—while stressing that the shift should be deliberate and not a panic‑driven scramble.
Why the Call for Bunker Mode?
Drake’s post followed OpenAI’s release of 722 new mathematical results the day before. He argued that these advances make it “reasonable to brace for the possibility that ECDSA breaks before Q‑day, in the worst case in months, not years.” The core of the concern is that if an attacker can recover private keys from exposed public keys, the entire account could be compromised.
He did not claim a break has already occurred; instead, he suggested that large holders consider moving most of their holdings to addresses that have never signed a transaction. The idea is simple: if a public key has never been used on‑chain, it has not been exposed to the same level of scrutiny and may be harder to target.
Mixed Reactions from the Community
The response was swift and polarized:
- Yehuda Lindell (Coinbase cryptography lead) called Drake’s post a “really bad take,” noting there is “no evidence whatsoever pointing to a break of decades‑old hardness assumptions like elliptic curve cryptography.”
- Charles Guillemet (Ledger CTO) described the scenario as “FUD” but acknowledged that avoiding address reuse is reasonable for large holders. He warned that a mass migration could create operational mistakes that outweigh the risk being mitigated.
- Dankrad Feist argued that if public keys become exploitable, “your coins are going to zero,” and bunker mode would not help.
- Mert Mumtaz (Helius) dismissed the advice as “unnecessarily performative.”
- Haseeb Qureshi (Dragonfly) supported the warning, calling it a sober reminder that “conventional mathematics overturning unproven cryptographic hardness assumptions” is a real risk.
Vitalik Buterin, co‑founder of Ethereum, took a cautious stance. He said he does not recommend a scramble to move funds today but agrees that the risk is serious. He added that keeping funds in unused addresses is a good idea if it is easy, and cautioned that he has lost more money in botched migrations than in all hacks combined.
Practical Takeaways for Holders
- Assess Your Exposure – If you hold a large amount of ETH or tokens, check whether your addresses have ever signed a transaction. If not, you may already be in a safer position.
- Plan, Don’t Panic – A rushed migration can lead to lost funds. If you decide to move, do so in a controlled, well‑tested manner.
- Explore Post‑Quantum Alternatives – Schemes like ML‑DSA and other lattice‑based signatures are gaining traction. Consider diversifying your key infrastructure.
- Stay Informed – Follow updates from reputable cryptographers and the Ethereum Foundation. The situation is evolving, and new research may change the risk profile.
Bottom Line
While the threat of AI‑accelerated math breaking ECDSA is still theoretical, the debate underscores the importance of vigilance and preparedness. Bunker mode is a precautionary measure that may be prudent for large holders, but it is not a silver bullet. A measured, informed approach—combined with a move toward post‑quantum cryptography—offers the best path forward.
---
TL;DR: Ethereum researcher Justin Drake warns that AI‑driven math could soon threaten ECDSA, urging a cautious shift to unused addresses, but the community remains divided on the urgency and practicality of such a move.
Source
Read original sourceWhy we picked this
Security research on AI threats to crypto keys, well-supported analysis.