Quantum annealing is an analog model of quantum computation that encodes a problem as the low-energy configuration of a physical Hamiltonian, then slowly evolves the system from an easy-to-prepare starting Hamiltonian toward that problem Hamiltonian. If the evolution is slow enough relative to the system’s energy gaps, the system stays close to its instantaneous ground state throughout, ending near the problem’s low-energy solution. This differs from the gate-based (circuit) model of quantum computing, which builds a computation out of a discrete sequence of controlled operations on qubits.

Quantum annealing hardware, such as D-Wave’s systems, is also used more broadly as a programmable analog simulator of quantum many-body dynamics rather than only as an optimizer — by preparing the system with a Hamiltonian that itself matches the physics under study, such as a disordered Ising spin glass, and reading out how it evolves over time.