Quantum computing has a noise problem that is easy to state and hard to solve. The qubits at the heart of a quantum computer are fragile: any disturbance from heat, vibration or stray electromagnetic fields can corrupt the information they hold. Building a machine that can correct its own mistakes is the central engineering challenge, and progress is measured in how well that correction works rather than in how many qubits a device claims.
Why qubits are fragile
A classical bit is either a zero or a one, and it tolerates a good deal of interference before it flips. A qubit exploits superposition and entanglement, which are precisely the properties that make it useful and delicate. The same sensitivity that allows quantum algorithms to work also makes the hardware vulnerable to decoherence, the loss of quantum state over time.
What error correction requires
Correcting a quantum error is not simply copying data, because copying a qubit is forbidden by the no-cloning theorem. Instead, information is spread across many physical qubits so that an error in one can be detected without measuring the logical value. The overhead is substantial: protecting one reliable logical qubit may require hundreds or thousands of physical ones.
Error correction only helps if the physical error rate is below a threshold. Above it, adding more qubits makes things worse; below it, adding more improves reliability. Demonstrating that a system operates below threshold has been a milestone for the field, because it shows the approach can in principle scale.
Different hardware, different problems
Superconducting circuits, trapped ions, neutral atoms and photonic systems each have distinct error profiles. A technique that works for one may not transfer to another. This diversity is a strength, because it means the field is not betting on a single approach, and it makes comparisons difficult, because different platforms measure different things.
The applications most often discussed are simulating molecules for chemistry and materials, and certain optimisation and cryptography problems. For most of these, it is not yet clear that a quantum machine will outperform the best classical methods, which are also improving. The honest position is that the advantage is likely to appear first in narrow, specialised problems.
The cryptography question
A large, fault-tolerant quantum computer could break some widely used public-key encryption, which is why standards bodies have been developing post-quantum algorithms. Migration takes years, and the data being encrypted today may need to remain secret for decades. This is one area where the response is proceeding before the threat fully exists.
Headlines about qubit counts are less informative than they appear. A device with many noisy qubits may be less capable than one with fewer, better ones. The useful questions are about error rates, connectivity, coherence times and whether a logical qubit has been demonstrated. Those are the measures that determine progress.
The talent and cost problem
Quantum computing requires specialists in physics, engineering and computer science, and they are scarce. It also requires expensive equipment: dilution refrigerators, lasers and specialised fabrication. The result is a field dominated by a small number of well-funded laboratories and companies. That concentration affects which approaches are pursued and which questions get asked.
Classical computing has not stood still. New algorithms and hardware continue to improve performance on problems once assumed to need quantum machines. This means the bar for quantum advantage keeps rising, and some proposed applications may be overtaken before quantum hardware is ready. The competition between the two is a defining feature of the field.
The timeline question
Predictions about useful quantum computing have been consistently optimistic. The honest position is that fault-tolerant machines capable of solving commercially relevant problems remain years away, and that the first applications are likely to be narrow. Investors and governments are placing bets, and the payoff is uncertain.
For most companies, the practical question is not when a quantum computer will arrive but which cryptographic systems need replacing and when. Post-quantum migration is a multi-year project, and organisations with long-lived sensitive data face the earliest deadlines. The quantum threat is real, distant and already shaping security planning.
Why the field still attracts investment
Despite the uncertainty, governments and companies continue to invest, because the potential applications in chemistry, materials and security are valuable enough to justify a long bet. Public funding also supports the research base and the skills pipeline. The investment is a wager on a possibility, and it is being made deliberately.
The most defensible expectation is that quantum computing will arrive gradually and unevenly, transforming a few narrow problems before it transforms anything else. Planning for that, rather than for a sudden revolution, is the practical response.
This report draws on National Institute of Standards and Technology (NIST) and IBM Quantum.
Image: NASA Jet Propulsion Laboratory · Public domain · via Wikimedia Commons.