Gate-Based Quantum Computing
Gate-based quantum computers are the systems most people picture when they think about quantum computing. IBM, Google, IonQ, Quantinuum, PsiQuantum, Rigetti, QuEra, Atom Computing, and a growing number of other companies are building them. They are also called circuit models or universal quantum computers, and those terms carry real meaning.
A gate-based quantum computer operates by applying a sequence of precisely controlled operations, called gates, to qubits. A Hadamard gate transforms a computational basis state into a superposition of basis states. A CNOT gate entangles two qubits. Rotation gates adjust the phase of a qubit’s quantum state. The sequence of gates constitutes a quantum circuit, and the circuit encodes the algorithm being run. The output is obtained by measuring the qubits at the end of the circuit.
The word “universal” means that gate-based quantum computers can in principle run any quantum algorithm. They are the quantum equivalent of a general-purpose classical computer. Grover’s algorithm for database search, Shor’s algorithm for factoring large numbers, quantum simulation of chemical and physical systems, variational quantum algorithms for optimization, all of these run on gate-based hardware. Programming frameworks such as Qiskit, developed by IBM, and Cirq, developed by Google, are designed for this model. They allow researchers and developers to write quantum circuits in software and execute them on real hardware through cloud access.
What Gate-Based Systems Can Do Today
Gate-based quantum computers are powerful in a narrow and carefully defined sense, and limited in ways that matter for practical applications.
On the capability side, gate-based systems have demonstrated quantum supremacy, meaning they have performed specific computational tasks faster than any known classical algorithm could on the best available classical hardware. Google’s 2019 demonstration with its Sycamore processor and more recent results from Google, Quantinuum, and QuEra have pushed this boundary further. Several hardware platforms have now demonstrated gate fidelities that exceed fault-tolerance thresholds, the point at which error correction becomes a net benefit rather than a net cost. This is an important prerequisite for scalable fault-tolerant quantum computing.
On the limitation side, the number of reliable logical qubits available on current hardware remains small. Running Shor’s algorithm against real-world RSA encryption would require millions of physical qubits to encode the thousands of error-corrected logical qubits the algorithm needs at cryptographic scale. Current systems contain hundreds to low thousands of physical qubits, but only a small number can presently be organized into error-corrected logical qubits. Decoherence, the tendency of quantum states to collapse due to interaction with the environment, limits how long a circuit can run before errors accumulate beyond recovery.
The practical applications available on current gate-based hardware fall into two categories. The first is research and algorithm development, where simulators and small real hardware runs are used to develop and test quantum algorithms that will matter more when hardware scales. The second is early quantum advantage demonstrations on narrow, carefully chosen problems where the quantum speedup is real but the practical value is still being established.
Many researchers believe quantum simulation of chemistry and materials science remains one of the most promising long-term applications, where even modest quantum advantage could accelerate drug discovery and materials design. Certain optimization problems, where quantum heuristics outperform classical alternatives on specific problem structures, represent another near-term candidate for practical advantage.
Quantum Annealing
Quantum annealing is a fundamentally different computational model. D-Wave Systems, founded in 1999, is the primary commercial developer of quantum annealers and has been selling access to them since 2011, making D-Wave the longest-running commercial quantum computing company by a significant margin.
A quantum annealer does not use gates. It does not execute circuits. It cannot run Grover’s algorithm or Shor’s algorithm. Instead, it solves a specific class of problems by exploiting a physical process called quantum tunneling.
The way it works is this. The problem to be solved is mapped onto an energy landscape, where the correct answer corresponds to the lowest energy state of the system, called the ground state. The annealer starts all its qubits in superposition and then slowly evolves the system, reducing quantum fluctuations over time, in a process that allows the system to tunnel through energy barriers rather than having to climb over them classically. The hope is that quantum tunneling allows the system to find lower energy valleys, meaning better solutions, faster than classical optimization methods that must navigate the landscape without tunneling.
D-Wave’s latest systems have thousands of qubits, far more than any current gate-based system. However, those qubits are specialized for the annealing process and are connected in a specific hardware graph that constrains which problems can be mapped efficiently onto the hardware. Not every optimization problem maps cleanly onto D-Wave’s qubit connectivity structure, and those that do not incur additional overhead that can reduce or eliminate any quantum advantage.
What Quantum Annealers Can Do Today
Quantum annealers are designed for combinatorial optimization problems: finding the best solution among an exponentially large number of possibilities. Real-world problems in this category include vehicle routing, supply chain optimization, financial portfolio optimization, protein folding, traffic flow management, and certain machine learning training problems. These are problems that classical computers find hard not because they are theoretically impossible but because the search space is too large to explore exhaustively and heuristic methods do not always find the best solution.
D-Wave’s systems are being used commercially today by companies including Volkswagen, Lockheed Martin, and various financial institutions for problems in this category. In many cases, these deployments are exploratory or hybrid workflows rather than demonstrated examples of broad quantum advantage.
The barrier to entry is lower than that of gate-based systems in some respects because D-Wave provides a relatively accessible programming model through its Ocean framework, and the problems it targets are ones businesses already care about solving.
The Contested Question
The central unresolved question about quantum annealing is whether it actually provides quantum speedup over the best available classical optimization algorithms. This is not a settled matter.
Some benchmarks show D-Wave performing well against classical alternatives on specific problem instances. Others show carefully tuned classical algorithms matching or outperforming D-Wave on the same problems. Simulated annealing, a classical algorithm that mimics the annealing process without quantum mechanics, is among the most competitive.
The difficulty is that the best classical optimization algorithms are themselves very good, and demonstrating a consistent, general quantum advantage over them on practical problem sizes has proven elusive.
This does not mean quantum annealing is not doing quantum computation. D-Wave’s hardware uses genuine superconducting qubits cooled to millikelvin temperatures and exploits real quantum-mechanical phenomena, including tunneling and entanglement. The debate is not about whether the physics is quantum. It is about whether that quantum physics translates into computational advantage over classical methods for the problems people actually want to solve.
Many researchers remain unconvinced that quantum annealers have demonstrated a broad and reproducible computational advantage over the best classical optimization methods on practical problems. While there is substantial evidence that quantum tunneling and other quantum effects play a role in annealing hardware, whether those effects translate into consistent performance advantages remains an active area of research.
How the Two Camps Compare
The two approaches are solving different problems in different ways and should not be thought of as competing for the same territory.
Gate-based quantum computers are universal. They can run any quantum algorithm in principle and are the platform on which the field’s most important theoretical results, Shor’s algorithm, Grover’s algorithm, and quantum simulation, are implemented. Their limitation is that error correction overhead is enormous and the hardware required for practically useful computation at scale does not yet exist, though the trajectory of recent hardware progress has become meaningfully more optimistic.
Quantum annealers are specialized. They target a specific and commercially important class of optimization problems and are available and in commercial use today. Their limitation is that evidence for quantum advantage over classical methods remains contested, and the range of problems they can address is narrower than that of gate-based systems.
A useful analogy is the difference between a general-purpose classical computer and a specialized signal processor. The general-purpose computer can run any program. The signal processor is faster at the specific tasks it was designed for, but it cannot run arbitrary programs. Gate-based quantum computers are the general-purpose systems. Quantum annealers are the specialized processors. Both have a role. Neither makes the other obsolete.
Where the Field Stands
Gate-based quantum computing is experiencing genuine, accelerating hardware progress. Qubit fidelity has crossed fault-tolerance thresholds on multiple platforms in the past two years. Several hardware companies have published results that represent meaningful steps toward practical quantum advantage. The consensus among researchers has shifted from guarded skepticism about near-term progress to cautious optimism about the next decade, a shift driven by hardware results rather than theoretical promises.
Quantum annealing is in commercial use today for optimization problems, which gives it a practical foothold that gate-based systems have not yet achieved at the same scale. Whether that foothold represents genuine quantum advantage or sophisticated classical-equivalent computation remains the field’s most pointed open question about the technology.
What both camps share is a trajectory. The hardware is getting better. The algorithms are getting more sophisticated. The engineering challenges are being solved one by one. For many researchers, the question has increasingly shifted from whether large-scale quantum advantage is possible to when and where it will first emerge.