Justin Drake calls for a 'bunker model' to deal with rapidly growing AI threats

Ethereum researcher Justin Drake is calling on the industry to enter a "bunker model", advising users to gradually move funds to new wallet addresses

Ethereum researcher Justin Drake is urging the cryptocurrency industry to start planning for what he calls a "bunker model" scenario and advising users to gradually move funds to new wallet addresses. His warning is based on recent advances in artificial intelligence and advanced mathematical problem solving capabilities that he believes could shorten the timeline for cracking the Elliptic Curve Digital Signature Algorithm (ECDSA), which is widely used to secure many cryptocurrency wallets.

In a post on Platform X (formerly Twitter) on Wednesday, Drake pointed out that advances in artificial intelligence-especially the work highlighted by OpenAI-are a reason to consider whether ECDSA is threatened earlier than many people assume, which in the worst case could be shortened from "years to months." Instead of waiting for a fully realized quantum computing breakthrough, he recommends preparing for cryptographic threats that may arrive earlier through other means.

Core Points

• Drake believes that recent mathematical advances related to artificial intelligence make it necessary to carefully consider that ECDSA security may degrade faster than expected.
• He recommends a "controlled" migration to a new wallet address rather than moving it all at once.
• New addresses can reduce exposure risks by hiding public keys behind cryptographic hashes until transactions reveal them again.
• Large, complex holders should prioritize the transfer of funds and subsequently process the remaining small balances after signing the transaction.
• Drake warned that a hasty migration could create additional operational risks that "would do more harm than good."

Why Drake is pushing for "bunker mode"

Drake's core concern is that artificial intelligence may eventually allow attackers to recover private keys from exposed public keys. Under his framework, the actual impact will be earlier than the point in time associated with quantum computing timetables in many discussions-because the associated cryptographic weaknesses may be achieved without waiting for a sufficiently powerful quantum machine.

He made clear that "there are reasons to prepare for the possibility of the ECDSA collapsing before the 'Quantum Day'," adding that in the worst case scenario, the time frame could be "measured in months rather than years." For users, this difference is crucial: It means that defensive measures should be considered immediately, rather than waiting until quantum capabilities become inevitable.

Drake also highlighted how difficult it has become to rely on inherent assumptions about mathematical intuitions for the long term. "The last few days have been a humbling lesson for human mathematical instincts. Long-standing and unquestioned assumptions have collapsed,"he wrote, linking this broader theme to breakthroughs he believed could have implications for cryptography.

OpenAI's mathematical progress as a signal

A key part of Drake's argument points to the development of artificial intelligence systems that seem to accelerate mathematical research. In a post citing OpenAI's work, he cited the company's report on sharing AI progress in mathematics, which described hundreds of new mathematical discoveries in fields including algebra, theoretical computer science and mathematical logic.

OpenAI's summary describes this as a recent meaningful leap forward in AI assistance. Drake further mentioned another update to OpenAI last month, using a large team of autonomous AI agents to solve the Navier-Stokes equation, which OpenAI said was completed in 88 hours, an example designed to demonstrate the speed and scale of solving problems.

In Drake's view, these advances suggest that cryptographic assumptions should be stress tested for the possibility that "new mathematics, not just new hardware, may change schedules." He also believes that elliptic curves may be particularly vulnerable to "fancy tricks," pointing out that the structured algebraic properties carried by curves may be exploited and that hashes are designed to minimize such structures.

Defensive move: Changing addresses

Drake's proposal is not an immediate abandonment of existing cryptography, but a practical mitigation: start moving funds to new addresses. "My personal recommendation is to initiate a large-scale controlled migration of assets to new locations," he wrote.

His method involves a phased plan. He advises larger and more mature holders to take the lead. He also advises users to transfer any remaining funds to a new address after signing a transaction. This distinction is important because Drake's reasoning focuses on the exposure of the public key.

According to Drake, certain types of addresses can effectively hide the public key behind a cryptographic hash. This is important because attacks to recover private keys from public keys often require knowing the public key. However, when a transaction is signed, the public key may become visible. As a result, he believes users should avoid leaving balances at addresses where their public key has been exposed.

At the same time, Drake emphasized that the migration must be gradual and carefully managed. He warned that "a hasty move can do more harm than good," which means that if mishandled, operational errors, coordination issues or avoidable transaction costs may outweigh the benefits.

What it means to go beyond a single wallet or a single chain

Although Drake's comments focused on ECDSA and address-level mitigation measures, their broader implications were about the speed with which cryptocurrencies respond to changing cryptographic risks. Many security discussions emphasize that quantum-driven risks require long preparation times; Drake is effectively urging the community to also plan for early failure modes brought about by advances in mathematics and AI reasoning.

For market participants, this guidance translates into a question of execution: How can large holders, institutions and ordinary users reduce cryptographic exposure without introducing new errors? Drake's emphasis on controlled relocation reflects this tension-security measures are only useful if they can be implemented safely and consistently.

Readers should pay attention to whether wallet providers, hosting companies, and developers incorporate such address rotation guidance into best practices, and whether there is more technical analysis clarifying what "monthly rather than annual" means for real-world wallet designs and threat models.