What Is a Quantum Error-Correcting Code? A Plain English Guide
Why do quantum computers need error correction?
Qubits are fragile. Heat, stray fields and imperfect control pulses disturb them, and errors pile up with every operation. A calculation that needs millions of operations, such as Shor's algorithm on a real key, cannot survive on raw qubits. John Preskill calls today's noisy machines NISQ devices (arXiv:1801.00862). Error correction is the answer. See why qubits are fragile.
How does classical error correction work?
An ordinary computer protects a bit by repeating it. Store 000 for a 0 and 111 for a 1. If one bit flips, a majority vote recovers the original. This repetition code fails for quantum information for three reasons. A quantum state cannot be copied (the no-cloning theorem). Reading a qubit to check it destroys its superposition. And quantum errors are not only bit flips: a qubit can also suffer phase errors and small continuous drifts.
The key idea: check parities, not values
Quantum codes avoid copying. They encode one logical qubit into an entangled state of several physical qubits. Then they measure syndromes, which are parity checks such as "are qubits 1 and 2 the same?". The answers reveal which qubit went wrong without revealing the encoded 0 or 1, so the superposition survives. A decoder, which is classical software, reads the syndrome pattern and chooses the correction. A simple example is the three qubit bit flip code: encode a qubit across three, check the parity of neighbouring pairs, and undo a single flip. It protects against one kind of error only. See logical vs physical qubits.
Famous codes
| Code | Idea |
|---|---|
| Shor code (1995) | The first quantum code. It uses nine qubits per logical qubit and protects against any single qubit error, bit flip or phase flip. |
| Steane code (1996) | Seven qubits per logical qubit, based on a classical Hamming code (arXiv:quant-ph/9601029). |
| Stabilizer codes (Gottesman, 1997) | A general framework that describes most codes through a set of parity checks (arXiv:quant-ph/9705052). |
| Surface code | Qubits on a 2D grid with local checks. It has a high error threshold and needs only nearby connections, which suits chips (Fowler et al., arXiv:1208.0928). |
| LDPC codes | Codes with sparse checks that need fewer physical qubits per logical qubit. IBM's "bivariate bicycle" family shows 12 logical qubits preserved with 288 physical qubits in total, assuming a 0.1% physical error rate (Bravyi et al., arXiv:2308.07915). |
| Cat qubit codes | Hardware that natively suppresses one kind of error so less code is needed on top. Amazon and Alice & Bob pursue this, and Amazon says its Ocelot chip can cut the cost by up to 90% (a company claim). |
What is the threshold?
Each code has an error threshold. If physical error rates are below it, adding more qubits to the code (a larger code distance) makes the logical error rate fall, often exponentially. If they are above it, bigger codes make things worse. The surface code's threshold is around 1%, and IBM's LDPC family reports 0.8% in its paper. Google's Willow chip, announced in December 2024, showed errors falling as the code grew, called below threshold, a necessary milestone for scaling (Nature, linked below). It is a demonstration on one logical qubit memory, not a full computer.
What does it cost?
Overhead is the main problem. A surface code logical qubit can need hundreds to thousands of physical qubits depending on how noisy the hardware is and how strong the protection must be, and a machine needs many logical qubits, which is why estimates for breaking cryptography run into hundreds of thousands of physical qubits. See how many qubits to break Bitcoin and IBM's roadmap to 200 logical qubits. Better codes, better hardware and smarter decoders, including AI decoders, all aim to shrink it. See quantum computing vs AI.
Does this matter for QNT?
Error correction decides when quantum computers become a real threat to cryptography. QUANTUM (QNT) is a community token on Solana and has no quantum hardware. See blockchain vs quantum computing.
Sources and further reading
- Preskill: Quantum Computing in the NISQ era and beyond (arXiv:1801.00862)
- Steane (1996): Multiple Particle Interference and Quantum Error Correction (arXiv:quant-ph/9601029)
- Gottesman (1997): Stabilizer Codes and Quantum Error Correction (arXiv:quant-ph/9705052)
- Fowler et al. (2012): Surface codes: Towards practical large-scale quantum computation (arXiv:1208.0928)
- Bravyi et al.: High-threshold and low-overhead fault-tolerant quantum memory (arXiv:2308.07915)
- Google: Meet Willow, our state-of-the-art quantum chip
- Nature: Quantum error correction below the surface code threshold
- Amazon: Ocelot quantum chip
- Alice & Bob: cat qubit computer for France (June 2026)
Checked 2026-10-12. Estimates are theoretical and company figures are the companies' own claims, so check the primary sources. This is education, not investment or security advice. QUANTUM (QNT) is an independent community token on Solana and is not affiliated with Quantinuum (Nasdaq: QNT) or Quant Network (QNT on Ethereum).
Frequently asked questions
What is a quantum error-correcting code?
A scheme that encodes one logical qubit across many physical qubits and uses parity checks to detect and fix errors without reading the data.
Why can't we just copy a qubit to protect it?
Quantum states cannot be copied (the no-cloning theorem), so codes use entanglement and parity checks instead.
What is the error threshold?
The physical error rate below which making a code bigger improves reliability. Above it, bigger codes make things worse.
Has error correction been demonstrated?
Yes in part. Google's Willow chip showed errors falling as the code grew in 2024, on a single logical qubit memory. Large error corrected computers are still being built.
Next read
Logical vs Physical Qubits: How Quantum Resource Estimates WorkWhy one paper says 1,450 qubits and another says under 500,000? Logical and physical qubits, gate counts, runtime and assumptions, explained plainly.Read nextAlso useful: Qubits Explained: What a Qubit Is, the Main Types and Why They Are Fragile | Google Quantum AI: Willow, Error Correction and Quantum Echoes | IBM's Quantum Roadmap: From Nighthawk to Starling in 2029
Educational content, not financial advice. QUANTUM (QNT) is independent of Quantinuum (Nasdaq: QNT) and Quant Network (QNT on Ethereum). Live chart | Token facts