SAXON Q and Fraunhofer IWU have published the first three-qubit Grover search executed on a commercial room-temperature quantum computer. The results, now on arXiv, show measured performance that exceeds every other published room-temperature quantum search and surpasses several cryogenic platforms on the same algorithm. In an eight-state search space, the system achieves 77.3% success probability for one marked state, compared to a classical ceiling of 37.5%. For two marked states, the figure rises to 87.0%, well above the classical limit of 46.4%. The only published result higher, 89.4% on a silicon spin qubit processor, required temperatures of 15 millikelvin and readout post-selection.

A commercial machine under standard conditions

The SAXON Q result was obtained on a commercial machine in a customer facility, at a measured average temperature of 296.3 Kelvin. It ran in open air with no vacuum chamber and no cryostat, using about 600 watts of total power. A dilution refrigerator and its supporting plant need tens of kilowatts. Researchers from SAXON Q in Leipzig and Fraunhofer IWU in Dresden operated a commercial NV-center diamond processor. The calculation qubits were three long-lived nuclear spins: the nitrogen-14 nucleus intrinsic to the defect and two nearby carbon-13 nuclei. The surrounding carbon lattice shields them, so they retain their quantum state for several milliseconds. That is roughly a thousand times longer than the NV center's own electron spin, which is used only as an optical read/write interface. Single-qubit gate fidelities reached as high as 99.98%, with an average single-qubit Clifford fidelity of 99.90%. Two-qubit subspace gate fidelities of a three-qubit gate averaged 95.7%.

Why Grover's algorithm matters

Grover's algorithm delivers a proven quadratic speedup over classical brute-force search. Running it end-to-end exercises a full stack of gates rather than a single isolated operation. That has made it a standard way of stress-testing what a quantum processor can actually do. Superconducting processors at millikelvin temperatures have reported comparable or somewhat lower success probabilities. Trapped-ion demonstrations of the same search reported 44 to 69%. They need no cryogenics but require sophisticated laser cooling and ultra-high vacuum. The diamond processor behaves like a solid-state ion trap. The NV centers and their nuclear spins are permanently fixed in a crystal lattice, so it needs no vacuum pumps, cryostat or optical traps. Read the full paper on Arxiv.org