Hardware & Errors
Fault Tolerance
The property of a quantum computer being able to perform reliable computation despite the presence of errors in individual components.
Definition
Fault tolerance describes a quantum computing system's ability to continue functioning correctly even when individual physical qubits or gates experience errors, by using quantum error correction to detect and fix problems faster than they accumulate.
In depth
Fault tolerance is the ability of a quantum computer to carry out arbitrarily long computations reliably even though its parts are imperfect. It goes beyond error correction in the narrow sense. A fault-tolerant design makes sure that a single error in a gate, a measurement, or an ancilla does not spread to cause a failure that the code cannot fix. That requires special constructions called gadgets for every operation, so that errors are caught before they multiply. The threshold theorem says that if the physical error rate is below a threshold, fault-tolerant protocols can suppress logical errors as much as needed.
A complete fault-tolerant machine needs several ingredients: encoded logical qubits, repeated syndrome measurement with fast classical decoding, a way to perform logical gates, including non-Clifford gates through magic states, and a way to prepare and measure logical qubits. Today's experiments show these ingredients one at a time. Google's below-threshold surface code memory, logical gate demonstrations on ions and atoms, and magic state distillation experiments are pieces of the puzzle. Microsoft describes three levels of progress: foundational noisy machines, resilient machines with logical qubits, and scale machines that solve problems classical computers cannot.
The cost is high. Estimates for useful algorithms need hundreds of thousands to millions of physical qubits, billions of operations, and decoders that keep up with microsecond cycles. Roadmaps from IBM, Google, Quantinuum, QuEra, and others target the late 2020s and early 2030s for the first fault-tolerant machines with tens to hundreds of logical qubits, and DARPA's Quantum Benchmarking Initiative aims to test whether a utility-scale machine is achievable by 2033.
Technical Definition
The fault-tolerance threshold theorem establishes that, as long as physical error rates are below a certain threshold, increasing the size of an error-correcting code (and the redundancy it provides) can suppress logical error rates arbitrarily, in principle enabling arbitrarily long, reliable computation.
Visual Explanation: An Analogy
Think of fault tolerance like a well-designed building with redundant support structures — even if a few support beams develop minor cracks, sufficient redundancy and reinforcement keep the building safely standing rather than collapsing.
Key facts
| Item | Detail |
|---|---|
| Definition | Reliable computation of any length despite imperfect gates, measurements, and preparation |
| Key theorem | The threshold theorem, proved in 1996 to 1998 |
| Microsoft's levels | Foundational, resilient, and scale |
| Typical needs | Hundreds of thousands to millions of physical qubits for useful algorithms |
| Targets | IBM Starling (2029) and DARPA QBI utility-scale test (2033) |
Where the idea comes from
Peter Shor's 1996 paper on fault-tolerant quantum computation and the threshold theorems of 1996 to 1998 established the field. Experimental progress accelerated in 2021 to 2025 with below-threshold results.
Real-World Use Cases
- The ultimate engineering goal of quantum hardware development, beyond today's NISQ-era machines
- Google's Willow processor's below-threshold error scaling result was a significant step toward demonstrating fault tolerance is achievable
How it differs from related ideas
- Error correction detects and fixes errors in stored information, while fault tolerance extends this to computations with faulty gates.
- A fault-tolerant machine is not error-free. Its logical error rate is low enough for the intended algorithm.
- Noisy intermediate-scale machines lack fault tolerance and rely on mitigation.
Frequently asked questions
What is fault-tolerant quantum computing?
A way of computing with error-corrected logical qubits and carefully designed operations so that long, reliable computations are possible despite noisy hardware.
When will fault-tolerant quantum computers exist?
Roadmaps target the late 2020s for early machines and the 2030s for ones that solve useful problems, though schedules are uncertain.
What is needed besides error-correcting codes?
Fault-tolerant gates, magic states for non-Clifford operations, fast decoders, and enough high-quality qubits.
Related Terms
Keep exploring
- Algorithm: Quantum Error-Correcting Codes (Shor Code, Surface Code)
- Algorithm: Shor's Algorithm
- Hardware: Google Willow
- Hardware: Quantinuum H2