Types of Quantum Computers: Superconducting, Ion, Photonic and More
Superconducting qubits
Tiny circuits cooled near absolute zero. Fast operations and chip-style manufacturing, but they need heavy refrigeration and are sensitive to noise. Used by groups such as IBM and Google.
Trapped ions
Charged atoms held by electric fields and controlled with lasers. Known for very accurate operations and good connectivity, but operations are typically slower. Companies working on this include IonQ and Quantinuum Ltd.
Photonic
Qubits encoded in particles of light on optical chips. They can work at room temperature for parts of the system and connect well to networking. Building the needed number of high-quality photons is hard. PsiQuantum is a well-known example.
Neutral atoms
Uncharged atoms arranged by laser beams ("optical tweezers"). They scale to large arrays of qubits, and companies such as QuEra and Pasqal use them.
Quantum annealers
A different design, used by D-Wave, aimed at optimization problems rather than general quantum computing.
Topological qubits
A research approach that aims for built-in error resistance. It remains an active and debated area.
Each of these still depends on error correction. For who builds what, see quantum computing companies.
Side by side
| Type | Examples | Strength | Open problem |
|---|---|---|---|
| Superconducting | Google Willow (105 qubits), IBM | Fast operations, chip fabrication | Cooling, wiring, noise |
| Trapped ion | Quantinuum Ltd Helios (98 qubits), IonQ | Accuracy, all-to-all links | Slower, scaling |
| Neutral atom | QuEra, Pasqal | Large atom arrays | Speed, atom loss |
| Photonic | PsiQuantum, Xanadu | Networking, less cooling in places | Photon loss |
| Annealer | D-Wave | Optimization focus | Not general purpose |
| Topological | Microsoft | Built-in protection (goal) | Evidence still debated |
What is changing in 2026
- Superconducting: IBM plans its Kookaburra module in 2026 en route to the Starling machine targeted for 2029. The 2025 Nobel Prize to Clarke, Devoret and Martinis highlighted the physics behind this approach. See superconducting qubits.
- Trapped ions: Helios launched in November 2025 with 98 qubits; the company quotes two-qubit gate fidelity of about 99.92 percent, a self-reported figure. See trapped ions.
- Neutral atoms: fast progress on large arrays. See neutral atoms and ions versus atoms.
- Topological: Microsoft announced its Majorana 1 chip in February 2025, and many physicists said the public evidence did not yet show a working topological qubit. See the Majorana debate.
Why annealers are different
D-Wave's annealers are a separate design. In March 2025 D-Wave claimed a result beyond classical reach on a magnetic simulation task, and other research groups published classical methods that matched parts of it, with D-Wave disputing the comparison. Read gate model versus annealing.
Common mistakes
- Declaring a winner. Different tasks may favor different designs.
- Comparing qubit counts across types. A hundred accurate, connected qubits can beat thousands of noisy ones.
- Treating a roadmap as a delivered machine.
Sources
- Google: Willow
- Quantum Computing Report: Helios
- Science News: physicists mostly unconvinced by Microsoft's chip
- Physics World: D-Wave advantage claim
- HPCwire: IBM roadmap
Frequently asked questions
Which type of quantum computer is best?
There is no agreed winner. Each approach has strengths, and the field is still competing.
What is a quantum annealer?
A specialised machine designed to find low-energy solutions to optimization problems. It is not a general quantum computer.
Which type will win?
Nobody knows. Several approaches may coexist, as different chip types do in classical computing.
Which types are available in the cloud?
Superconducting, trapped-ion, neutral-atom and annealing machines all have some cloud access. See free tiers compared.
What is a silicon spin qubit?
A qubit stored in the spin of an electron in a silicon chip, which could use existing chip factories. See silicon spin qubits.
Are photonic computers room temperature?
Parts can be, but detectors often need cooling. Read photonic quantum computing.
Keep reading
- What Is a Qubit? Superposition and Measurement Explained
A qubit is the basic unit of a quantum computer. Learn how qubits work, how they are built, and why they are fragile. - Quantum Computing Companies to Know (and Not to Confuse)
A neutral overview of well-known quantum computing companies and what approach each takes, plus how they differ from a memecoin. - Quantum Error Correction Explained
Qubits are fragile, so quantum computers need error correction. Learn how logical qubits are built and why this is the key challenge. - Superconducting Qubits Explained in Depth
How do superconducting qubits work? A clear look at circuits, microwave control, strengths, weaknesses and open questions in widely used quantum hardware. - Trapped Ion Quantum Computers Explained
How do trapped ion quantum computers work? Learn how charged atoms become qubits, why they are accurate, why they are slow, and how they might scale. - Neutral Atom Quantum Computers Explained
How do neutral atom quantum computers work? Learn about optical tweezers, Rydberg interactions, flexible layouts and the challenges this approach faces.
All Quantum computing guides | Back to top | Search the site
Main pages: Quantum computing explained | Quantum and crypto | Companies | Quantum news | Glossary