AI Could Break Encryption Before Quantum Computers, Ethereum Researcher Warns

Ethereum Researcher Says AI May Break Encryption Before Quantum Does

Ethereum Foundation researcher Justin Drake, long known for warning that quantum computers could eventually undermine crypto security, is now urging the blockchain industry to prepare for a different near-term risk: artificial intelligence.

In a Wednesday post on X, Drake called for the industry to calmly begin planning for what he described as “bunker mode.” His suggested precaution is a controlled, gradual migration of funds to fresh addresses that have never signed a transaction. Those addresses keep their public keys unpublished—effectively hidden behind a hash—until they are used.

The concern centers on ECDSA (Elliptic Curve Digital Signature Algorithm), the signature scheme used by both Bitcoin and Ethereum to authorize transactions. Drake said it is now reasonable to plan for a scenario where ECDSA could fail before the arrival of large-scale cryptographically relevant quantum computers—potentially “in months not years” in a worst case—driven by rapid AI progress in mathematics.

Drake framed “breaking” ECDSA as the ability to recover a wallet’s private key from public information in roughly a week using a large cluster of graphics cards. He emphasized that holders should not panic or rush, and he said he will push to accelerate Ethereum’s transition toward hash-based cryptography.

Drake said his thinking shifted based on how quickly AI systems are producing and solving advanced mathematical work. He pointed to 722 mathematical results OpenAI published on Tuesday, arguing that elliptic-curve cryptography may contain the kind of rich mathematical structure a highly capable system could exploit. By contrast, he described hashing—used in address derivation and other parts of crypto systems—as designed to resist pattern-based attacks.

The warning arrives amid ongoing work inside the Ethereum ecosystem to prepare for post-quantum security. In January 2026, the Ethereum Foundation formed a dedicated post-quantum team led by Thomas Corthier. It also established the Poseidon Prize, offering $1 billion for improvements in hash-based cryptographic primitives.

Quantum risk remains a central part of the broader context. In March 2026, researchers at Google Quantum AI published a paper concluding that breaking Ethereum’s cryptography could require about 1,200 logical qubits—far fewer than earlier estimates of “tens of thousands.” Drake cited that result as evidence the quantum threat has moved from purely theoretical to something worth concrete planning.

Others in the Ethereum research community have described the transition as already underway. In a separate interview reported by Cointelegraph, Ethereum Foundation cryptography researcher Antonio Sanso said Ethereum is already part of the way toward quantum resilience and expressed confidence the network can become quantum secure well before an attack is feasible. Still, he noted that timelines are uncertain, and past technology shifts—such as the rapid emergence of large language models and advances in ZK-proofs—have sometimes arrived earlier than expected.

Drake’s “bunker mode” recommendation focuses on a practical reality of many current crypto systems: in some wallet designs, public keys become visible once an address signs a transaction. Moving funds to unused addresses delays that exposure, reducing the surface area for any future attack that depends on public-key data.

• What happened: Justin Drake urged the industry to plan for “bunker mode” and migrate funds gradually to unused addresses whose public keys have not been revealed.
• Why it matters: Drake argues AI-driven math progress could put widely used signature schemes like ECDSA at risk sooner than expected, independent of quantum timelines.
• Broader context: Ethereum has active post-quantum initiatives, and recent research has lowered some estimates of what quantum computers might need to threaten today’s cryptography.