Quantum risk to crypto surfaces
New Google‑backed quantum research warns several attack paths could let quantum machines break current crypto primitives, potentially exposing ~$100bn on Ethereum and stressing Bitcoin’s post‑quantum roadmaps. That raises a tangible timeline question for wallet, exchange and custody risk models. (ambcrypto.com; coindesk.com)
Google Quantum AI published a 57‑page whitepaper on March 31, 2026 that lists Ethereum researcher Justin Drake and Stanford cryptographer Dan Boneh as contributors to the study. (coinmarketcap.com) The team provides two Shor‑algorithm circuit designs to break ECDLP‑256: one needing fewer than 1,200 logical qubits and ~90 million Toffoli gates, and a second using fewer than 1,450 logical qubits and ~70 million Toffoli gates. (research.google) The paper flags an “on‑spend” attack window as short as 9–12 minutes and models a quantum machine that recovers one private key every nine minutes, allowing it to sweep 1,000 exposed accounts in under nine days. (ambcrypto.com) Google’s analysis identifies concentrated exposures on Ethereum: the top 1,000 addresses hold roughly 20.5 million ETH, about 70 admin keys control ~2.5 million ETH (including stablecoin minting privileges), and major Layer‑2s and bridges hold another ~15 million ETH. (coinmarketcap.com) The paper warns the KZG trusted‑setup secret used for data‑availability proofs could be recovered from existing public data, creating a permanently exploitable classical forgery once recovered. (coinmarketcap.com) Google reports its updated estimates are roughly a 20× reduction in resources needed to break ECDLP‑256 versus earlier work, reiterates that no cryptographically‑relevant quantum computer exists today, sets a 2029 internal migration target to post‑quantum cryptography, and Justin Drake has said he assigns at least a 10% probability of a key‑recovery “Q‑day” by 2032. (bloomberg.com)
