New IonQ paper shows that wallets could be broken with only 20,000 trapped ion qubits, projected to be achieved before 2030

A new paper from IonQ shows that their trapped-ion quantum architecture can be optimized to run Shor's algorithm against 256-bit ECDSA signatures (which underlie most cryptocurrencies) with only 20,000 qubits in less than a month.

Part of it is algorithmic improvements (more efficient circuit to run the algorithm) but most of the *significant* drop in the qubit estimate comes from their new architecture, which requires only approximately 5 physical qubits per error-corrected logical qubit.

They are expected to release a 10,000-qubit trapped-ion quantum computer next year, with a 20,000 one coming before 2030. This is moving the timeframe forward by a significant amount...

The actual paper with the technical details is linked from this article. I tried to link the paper directly but it was flagged as an URL shortener falsely.