On 30 July 2026 IBM and the University of Chicago announced a computation encoding 70 logical qubits with a new error-correction method, completing in roughly 15 minutes a task they argue is beyond leading classical simulation, with verification that the result is trustworthy.
The significant part for cryptography is not the runtime but the combination of logical qubits and verification: progress in error correction is the variable that governs when cryptographically relevant machines become possible.
Published resource estimates for breaking RSA-2048 still run to millions of physical qubits or, in more recent optimised estimates, hundreds of thousands, operating for days. Seventy logical qubits does not approach that regime.