Level 4 · Hardware & Error Correction
Quantum Hardware
Building a quantum computer is one of the hardest engineering challenges in modern science. There's no single "right" way to build a qubit — instead, several fundamentally different physical approaches are being pursued in parallel, each with its own trade-offs. This article tours the major approaches and the key metrics used to compare them.
Superconducting qubits
Used by companies including IBM, Google, and Rigetti, superconducting qubits are tiny circuits made from superconducting materials — materials that conduct electricity with zero resistance below a certain temperature. These circuits are cooled to temperatures colder than outer space (a few thousandths of a degree above absolute zero) inside devices called dilution refrigerators.
Advantages: Superconducting qubits can be manufactured using techniques adapted from the semiconductor industry, and gate operations are extremely fast (nanoseconds). Challenges: They require extreme cooling infrastructure, and coherence times (how long a qubit maintains its state) are relatively short compared to some other approaches.
Trapped-ion qubits
Companies like IonQ and Quantinuum use trapped-ion qubits — individual atoms that have been stripped of an electron (making them charged "ions"), held in place by electromagnetic fields, and manipulated using precisely tuned lasers.
Advantages: Trapped ions are naturally identical to each other (every ion of the same element is the same), have long coherence times, and tend to have high gate fidelity (accuracy). Challenges: Operations involving lasers are generally slower than electrical pulses, and scaling up to large numbers of ions while maintaining precise control is a significant engineering challenge.
Photonic qubits
Photonic quantum computers encode information in properties of particles of light (photons), such as polarization.
Advantages: Photons don't require the extreme cooling that superconducting qubits do, and they're a natural fit for quantum communication and networking, since light is already how information travels over long distances (fiber optic cables). Challenges: Making photons reliably interact with each other — necessary for multi-qubit gates — is technically difficult.
Neutral atom qubits
Neutral atom approaches use arrays of uncharged atoms, held in place and arranged into custom patterns using laser "tweezers" (focused beams of light that can trap and move individual atoms).
Advantages: Neutral atom arrays have shown promise for scaling to large numbers of qubits, since the laser-tweezer approach can arrange many atoms in flexible, reconfigurable patterns. Challenges: Like trapped ions, operations rely on precise laser control, and the field is relatively newer than superconducting and trapped-ion approaches.
Topological qubits
Pursued primarily by Microsoft, topological qubits are a more experimental approach that aims to encode quantum information in a way that's inherently resistant to certain types of errors, based on exotic physical phenomena predicted by theory but requiring extremely precise materials science to realize.
Advantages: If realized successfully, topological qubits could dramatically reduce the error-correction overhead discussed in the previous article. Challenges: The underlying physical phenomena are extremely difficult to create and control reliably, and the approach remains earlier-stage than other technologies.
Key metrics for comparing quantum hardware
When comparing different quantum computers — whether different technologies or different generations of the same technology — a few key metrics matter more than raw qubit count alone:
| Metric | What it measures |
|---|---|
| Qubit count | How many physical qubits are on the chip or in the device |
| Coherence time | How long a qubit maintains its quantum state before decohering |
| Gate fidelity | How accurately a gate operation performs its intended transformation |
| Connectivity | Which qubits can directly interact with which others |
| Gate speed | How quickly operations can be performed |
A chip with a huge number of qubits but poor coherence times and low gate fidelity may, in practice, be less useful than a smaller chip with much higher quality — which is why "qubit count" alone is an incomplete (and sometimes misleading) headline number.
What's next?
You've now completed the core Learning Center — from your first qubit through algorithms and hardware. From here, explore the Companies Database to see how these technologies map onto real organizations, or revisit the Complete Guide for a high-level recap of everything.
Frequently Asked Questions
Which qubit technology is "winning"?
There's no clear winner yet, and it's possible different technologies will end up best-suited for different applications (for example, photonic qubits for networking, superconducting or trapped-ion qubits for computation). Major companies are pursuing different bets, and the field is still evolving rapidly.
Why do quantum computers need to be so cold?
For superconducting qubits specifically, extreme cold is required both for the superconducting effect itself to occur, and to minimize thermal noise that would otherwise disturb the delicate quantum states. Other technologies (like trapped ions and photonics) have different — sometimes less extreme — environmental requirements.
Can I build or buy a quantum computer myself?
Building one from scratch requires highly specialized facilities and expertise. However, several companies offer cloud access to real quantum hardware, allowing anyone to write and run small quantum programs over the internet without owning the physical device.