The phrase "quantum computer" covers two different machines. A gate-based quantum computer runs circuits of quantum gates, like the Qiskit and Cirq code this site teaches. A quantum annealer runs no circuits at all. It holds thousands of qubits and slowly cools them to find a low-energy state. D-Wave builds annealers, and the two families solve different problems.
The gate model: universal but demanding
A gate-based computer applies discrete quantum gates to a small number of qubits, building a quantum circuit. Any quantum algorithm, Shor's factoring, Grover's search, VQE chemistry, fits this model. The machine is universal, in principle capable of running any quantum computation.
The cost is error correction. Gate machines need fidelity high enough that hundreds of gates work before noise destroys the state, and full fault tolerance needs many physical qubits per logical qubit. See Understanding Quantum Error Correction.
The annealer: one problem, thousands of qubits
A quantum annealer starts its qubits in an easy ground state and slowly changes the system until it represents a target problem. Quantum mechanics lets the system tunnel through energy barriers to settle in a low-energy state, which encodes a good solution.
The catch: an annealer solves only optimization and sampling problems that map onto this energy-surface form. It cannot run Shor's algorithm or a general circuit. It is a specialized optimizer, not a general computer. D-Wave's machines hold over 5,000 qubits, far more than any gate machine, but those qubits do a narrower job.
What each one is good at
Annealing targets combinatorial optimization: routing, scheduling, portfolio balancing, and graph problems written as a QUBO. D-Wave's work with Nasdaq Verafin on financial crime detection sits in this lane.
Gate machines target the wider set: factoring, search, chemistry simulation, and optimization through QAOA. The QAOA tutorial and quantum annealing tutorial both solve Max-Cut, which makes the contrast concrete.
No proven speedup, different bets
Neither family has demonstrated a practical advantage over classical solvers on a commercial problem as of 2026. The annealer offers scale today, thousands of qubits solving optimization problems now. The gate machine offers generality later, once error correction lands. A telling data point: D-Wave itself published a gate-model result in Nature, hedging toward universal hardware while still selling annealers.
Which one to use
Pick the annealer when your problem is pure combinatorial optimization and you want scale today. Pick the gate machine when your problem is not optimization, or when you are building toward the fault-tolerant era. For learning, both are reachable: D-Wave's Ocean SDK for annealing, and free Qiskit simulators for the gate model.
- Quantum Annealing Explained: Solving Max-Cut with D-Wave's Ocean SDK
- QAOA Explained: Solving Max-Cut Step by Step
- Types of Quantum Computers Compared: where the annealer fits among the seven modalities