Quantum volume (QV) is a benchmarking metric introduced by IBM to provide a more meaningful measure of quantum computer performance than raw qubit count alone. A processor with 100 noisy, poorly connected qubits may be less capable than one with 50 higher-quality qubits. Quantum volume captures this by testing how large a square random circuit (equal depth and width) a quantum system can execute with acceptable fidelity. The metric is expressed as a power of 2 — a quantum volume of 2^n means the system can reliably run random circuits on n qubits with n layers of gates.
Leading quantum computing companies regularly report quantum volume improvements. Quantinuum has achieved some of the highest published quantum volumes, exceeding 2^20, reflecting the high fidelity of its trapped-ion hardware. IBM has steadily increased the quantum volume of its superconducting systems with each processor generation. IonQ has also used quantum volume as a key competitive benchmark.
While quantum volume has been valuable for tracking hardware progress, it has limitations as a performance metric. It does not capture performance on specific algorithms, does not account for classical processing overhead, and becomes less meaningful as quantum computers grow beyond the NISQ era. Alternative benchmarks like algorithmic qubits (proposed by Google), CLOPS (circuit layer operations per second, from IBM), and application-specific benchmarks are emerging to complement quantum volume as the field matures. For deeper coverage, see DeepTechIntel's quantum computing section.