On September 23, 2026, the EU's three financial supervisors (banking, insurance and securities regulators, formally the EBA, EIOPA and ESMA and jointly the "ESAs") published their autumn 2026 risk report, warning that quantum computing could undermine the cryptography securing "communications, transactions, databases, and blockchains," and that the threat could arrive "earlier than any viable commercial application" because stolen encrypted data can be decrypted later, security researchers call this harvest now, decrypt later. For bitcoin specifically, the report cites Glassnode's estimate that 6.04 million BTC (30.2% of supply) sit in addresses with a public key (the value a signature can be checked against) already visible on-chain, the precondition a quantum computer would need to derive the matching private key. Independent estimates from CryptoQuant, Google Quantum AI and Coinbase put the exposed figure closer to 6.9 million once reused and Taproot addresses are counted; CoinDesk's reporting on the Google paper adds that deriving a key this way would take "roughly nine minutes" — fast enough to threaten a transaction still sitting unconfirmed in the mempool (the pool of transactions waiting to be added to a block), not only coins untouched for years.

This lands on Security, Decentralization & Durability: 18 of the framework's 100 points, which covers audit quality and a track record through at least one full market cycle. It's an unusual entry for that dimension: not one project's governance or node count, but the durability of the elliptic-curve signature math nearly every covered asset still relies on.

It's also a gap, not a flag. None of the nine hard gates (structural failures that force a red verdict regardless of score) name this risk. Gate 2 (unaudited or previously exploited contracts) checks whether code has been reviewed and whether it has actually been broken; it has no line for a primitive that remains unbroken today and might not be tomorrow. That's not a defect so much as a reminder of what gates are built to do: catch failures that already happened. A risk with no incident, no attacker, and (per Google's own estimate) a hardware gap (roughly 500,000 physical qubits needed, against roughly 1,000 in today's largest processors) that isn't close, doesn't trip a gate built for rugs and exploits. It is the kind of input the durability score would need to weigh if that gap ever started closing quickly. It hasn't yet.

One dry note: Taproot, Bitcoin's 2021 upgrade, was sold on privacy and efficiency gains. It is also the one change in this story that made exposure larger rather than smaller, a reminder that "forward" and "forward-secure" are not always the same direction.

CryptoGrade publishes educational research, not investment advice. Ratings and commentary are produced by a rules-based framework applied identically to every asset. Crypto assets are volatile and you can lose your entire investment. Do your own research.