D-Wave says a new building-block gate could make fault-tolerant quantum computers cheaper to build by trimming the number of physical qubits needed for error correction.
The gate, described this week in the journal Nature, works with the company’s superconducting dual-rail qubits. Each qubit is encoded across pairs of microwave photons, so a failure announces itself instead of silently corrupting a calculation. D-Wave reports roughly 99.9 percent fidelity on two-qubit operations, with gate times near 500 nanoseconds.
Simulations show the architecture could cut logical error rates by as much as 10x with each step of error correction, a property the company calls a favorable error hierarchy. That translates into far fewer spare qubits devoted to keeping a computation honest.
The roadmap leans on the design. A 49-physical-qubit system due next year should cut physical error rates 20-fold, and a 181-physical-qubit machine the year after targets a 2,000-fold reduction, which D-Wave says will set the blueprint for scalable fault-tolerant architectures.
Rob Schoelkopf, co-founder of Quantum Circuits and now chief scientist at D-Wave, has argued the dual-rail qubit’s built-in error detection gives it an edge over rival approaches. The company is aiming for a tenfold error-reduction factor per correction layer, a benchmark it calls Lambda 10.
