D-Wave Quantum unveiled its roadmap for the phased development of quantum systems. By 2032, the company plans to create a fault-tolerant quantum computer with 100 logical qubits.
According to the press release, the system is expected to perform over 1 million operations and support initial applications in quantum chemistry and AI.
Intermediate steps in D-Wave’s plan include:
- 2026: A system with 17 physical qubits and a logical error rate half that of classical systems;
- 2027: A 49-qubit system with an anticipated 20-fold reduction in errors;
- 2028: A 181-qubit system with an anticipated 2000-fold reduction in errors;
- 2030: A system with 10 logical qubits designed for initial fault-tolerant algorithms.
The company’s approach is based on a superconducting dual-rail qubit architecture, enabling quantum error correction cycles 100-1000 times faster than systems based on neutral atoms or trapped ions.
The company claims that fault detection is directly integrated into the qubits, allowing for the identification of about 90% of computational inaccuracies at the initial level. This solution significantly reduces the number of physical qubits needed for correction.
“The future of commercial gate-model quantum computing will be defined not just by the number of physical qubits, but by the ability to reliably perform large-scale computations for real-world applications. While much of the industry focuses on scaling physical qubits, D-Wave employs a differentiated approach focused on reducing hardware-level errors,” the company noted.
D-Wave also announced achieving 99.9% accuracy for two-qubit operations, with one physical error occurring per 1000 operations.
The company highlighted the Lambda metric, which describes how quickly error rates decrease with added correction. D-Wave estimates the industry average at about 2, while their plan targets 10.
In May, experts questioned D-Wave’s claims of quantum supremacy.