Superconducting Quantum Computers
Superconducting qubits are the most widely deployed quantum computing platform. Artificial atoms are constructed from superconducting circuits cooled to millikelvin temperatures, where quantum effects dominate. The transmon qubit—a charge-insensitive variant—is the dominant design. IBM, Google, and Rigetti lead commercial deployment; IBM Quantum offers cloud access to dozens of processors.
GATE SPEED
Single-qubit gates: ~10–100 ns
CONNECTIVITY
Fixed, nearest-neighbor topology (e.g., …
ENVIRONMENT
~15 millikelvin — dilution refrigerator required
How Superconducting Qubits Work
Superconducting qubits are artificial atoms built from electrical circuits. At the heart of each qubit is a Josephson junction — a thin insulating barrier sandwiched between two superconducting electrodes. When cooled to millikelvin temperatures, quantum mechanical tunneling through the junction creates a nonlinear inductance that discretizes the circuit's energy levels. The lowest two levels are used as the qubit's |0⟩ and |1⟩ states. The transmon variant, which suppresses charge noise by operating at high Josephson-to-charging energy ratios, is now the dominant design across IBM, Google, and Rigetti systems.
Quantum gates are implemented by applying precisely shaped microwave pulses through coaxial lines connected to each qubit's control port. Single-qubit gates rotate the qubit's state vector on the Bloch sphere; two-qubit gates like the CNOT or CZ gate are implemented by temporarily coupling neighboring qubits through a tunable coupler or cross-resonance interaction. Readout is accomplished by coupling each qubit to a dedicated readout resonator: the qubit's state shifts the resonator's frequency, which is detected by a heterodyne measurement chain at room temperature.
Fabrication uses standard semiconductor techniques — electron beam lithography and shadow evaporation — to deposit aluminum or niobium thin films on silicon or sapphire substrates. This compatibility with existing semiconductor manufacturing infrastructure is a significant advantage for scaling, though it also means qubit frequencies must be carefully engineered to avoid spectator modes and crosstalk.
Physical Implementation
Connectivity and Gate Operations
Connectivity
Fixed, nearest-neighbor topology defined at fabrication time. IBM uses the heavy-hex lattice (each qubit connected to 2–3 neighbors); Google uses a square lattice. Long-range connectivity requires SWAP networks with significant gate overhead. No reconfiguration is possible during runtime.
Native Gate Set
Single-qubit: arbitrary SU(2) rotation via Euler decomposition (Rx, Ry, Rz, X, Y, Z, H). Two-qubit: CX (CNOT), CZ, ECR (echoed cross-resonance), or iSWAP depending on architecture. Virtual Z gates are implemented as phase offsets at zero cost.
Gate Speed
Single-qubit gates: ~10–100 ns; two-qubit gates: ~40–500 ns
Potential Advantages
- ✓Fast gate operations — single-qubit gates in 10–50 ns, two-qubit gates in 40–500 ns — enabling high circuit throughput
- ✓Mature, semiconductor-compatible fabrication allows rapid iteration and scaling via established industry processes
- ✓Mid-circuit measurement and reset enable dynamic circuits and quantum error correction protocols
- ✓Largest commercial ecosystem: IBM Quantum, Google Quantum AI, Rigetti, IQM offer cloud access to dozens of processors
- ✓Microwave control at GHz frequencies is directly compatible with high-speed classical electronics
- ✓Demonstrated quantum error correction with distance-5 and distance-7 surface codes (Google "Willow" 2024)
Limitations & Challenges
- —Short coherence times (T1/T2 typically 50–500 µs) limit the number of gates executable before decoherence
- —Fixed nearest-neighbor connectivity requires SWAP overhead for algorithms expecting all-to-all access
- —Dilution refrigerators are expensive (~$1–2M), require weeks to cool, and limit qubit count by cooling power and wiring density
- —Frequency crowding: as qubit count grows, manufacturing identical-frequency qubits becomes increasingly difficult without crosstalk
- —Two-qubit gate fidelity of 99.0–99.9% still requires significant error correction overhead for fault-tolerant operation
- —Qubit uniformity varies across chips; not all qubits on a processor perform equally well
Scaling Considerations
IBM's roadmap targets modular quantum systems interconnected via high-bandwidth classical links and eventually quantum links. Google pursues increasing chip size with heavy-hex and similar layouts. The primary scaling challenge is maintaining coherence and fidelity as qubit count grows — crosstalk, frequency crowding, and wiring density all worsen. Error correction via surface codes is the most studied path to logical qubits; current estimates suggest ~1,000 physical qubits per logical qubit at surface code distance 25, though this ratio may improve with better fabrication. Google's Willow chip (2024) demonstrated below-threshold error correction where adding more qubits improved logical qubit performance — a critical milestone.
Coherence: T1 (relaxation): typically 50–500 µs; T2 (dephasing): 10–300 µs. Short coherence times compared to trapped-ion but compensated by fast gates.
Available QPUs
View all →Leading Companies
Workload Suitability
Full guide →Simulating molecular electronic structure to compute ground-state energies, reac…
Simulating solid-state physics phenomena including high-temperature superconduct…
Solving NP-hard combinatorial problems including vehicle routing, supply chain s…
Monte Carlo simulation acceleration via quantum amplitude estimation, derivative…
Training and inference on quantum neural networks (QNNs), kernel methods using q…
Shor's algorithm can factor large integers and solve discrete logarithms exponen…
WELL SUITED FOR
- ›Quantum chemistry (VQE, QSCI) with moderate circuit depth
- ›Variational quantum algorithms (QAOA, VQE) on near-term hardware
- ›Quantum error correction research and demonstration
- ›Machine learning circuits (QNNs, kernel methods)
- ›Randomized benchmarking and hardware characterization
LESS SUITED FOR
- —Algorithms requiring deep circuits or many serial two-qubit gates (decoherence limits)
- —Problems requiring all-to-all connectivity without SWAP overhead budget
- —Workloads benefiting from long coherence times between measurements
Frequently Asked Questions
Key References
- [1]Quantum supremacy using a programmable superconducting processor — Nature / Google, 2019
- [2]Quantum error correction below the surface code threshold — Nature / Google, 2024
- [3]Circuit quantum electrodynamics — Reviews of Modern Physics, 2021
- [4]IBM Quantum System Two — IBM, 2023