Quantum Computing Without the Hype

Quantum computing is often described as the next leap in computing power, a machine that will make today's supercomputers obsolete. The reality is narrower and, in some ways, more interesting. Quantum computers are not faster versions of ordinary computers. They are a different kind of machine that should be dramatically better at a small set of problems and no better, or worse, at almost everything else.

What makes them different

An ordinary computer stores information in bits, each either 0 or 1. A quantum computer uses qubits, which can be in a combination of 0 and 1 at the same time, a property called superposition. Qubits can also be entangled, so that the state of one is linked to the state of another. Quantum algorithms use these properties, together with interference, to steer a calculation toward the right answer and away from the wrong ones.

It is not accurate to say a quantum computer tries every answer at once and picks the best. For most problems that trick does not work. The speedups come from specific algorithms that exploit the structure of particular problems.

The problems they could help with

Simulating molecules and materials. Chemistry is governed by quantum physics, so ordinary computers struggle to simulate even modest molecules exactly. This was the original motivation, suggested by physicist Richard Feynman in the early 1980s, and it remains one of the most promising uses, with possible applications in drug discovery, batteries and fertilizer production.

Breaking some encryption. In 1994 mathematician Peter Shor showed that a large enough quantum computer could factor big numbers efficiently, which would break the RSA and elliptic-curve encryption that protects much of the internet. No existing machine is close to large enough, but the threat is taken seriously.

Some optimization and search problems. Here the expected gains are smaller and less certain, and claims of near-term business advantage deserve some skepticism.

Where the technology stands

In 2019 Google reported that its 53-qubit Sycamore processor completed a specially designed sampling task far faster than a classical supercomputer could, a milestone it called quantum supremacy. The task had no practical use, and classical researchers later found faster ways to simulate it, but it showed the machines could do something hard.

Since then, the focus has shifted to errors. Qubits are fragile. Heat, vibration and stray electromagnetic signals disturb them, and today's machines make mistakes far too often for long calculations. The solution is error correction: combining many physical qubits into one more reliable "logical" qubit. In late 2024 Google reported, with its Willow chip, that adding more qubits to its error-correcting code made errors fall rather than rise, a threshold researchers had been working toward for decades. IBM, Quantinuum, IonQ and others are pursuing their own approaches, using superconducting circuits, trapped ions or neutral atoms.

The competing hardware approaches have different trade-offs. Superconducting qubits, used by Google and IBM, are fast but must be cooled to a fraction of a degree above absolute zero. Trapped ions, used by Quantinuum and IonQ, are slower but tend to be more accurate and better connected. Neutral atoms held by lasers can be packed in large numbers. No one yet knows which will scale best.

Useful, fault-tolerant machines are likely to need many thousands of high-quality qubits or more. Estimates of when that will happen vary widely, and timelines in this field have often slipped.

Why encryption is changing now

Even though code-breaking machines don't exist yet, there is a concern known as "harvest now, decrypt later": data intercepted today could be stored and decrypted years from now. In August 2024 the US National Institute of Standards and Technology published its first post-quantum cryptography standards, encryption methods designed to resist quantum attacks. Browsers, messaging apps and operating systems have begun adopting them, usually without users noticing anything.

Common misconceptions

"More qubits means more power." Qubit counts make headlines, but quality matters as much as quantity. IBM built a processor with more than 1,000 qubits in 2023, yet the error rate of each qubit and how well they are connected determine what a machine can actually do.

"Quantum computers will replace laptops." They are specialized machines. Even in the most optimistic scenarios, they would work alongside ordinary computers, handling particular parts of particular problems.

"Quantum advantage has already arrived for business." Companies and research groups regularly announce experiments, but a clear, practical case where a quantum computer beats the best classical methods on a problem someone actually needs solved has been hard to demonstrate. When a claim appears, it is worth checking whether classical researchers have tried to match it.

What it means for everyday life

For now, very little directly. Quantum computers are large, kept at temperatures close to absolute zero in many designs, and accessed by researchers and companies through the cloud. You will not have one on your desk. The near-term effects are more likely to be indirect: quieter changes in the encryption your devices use, and, if the hardware matures, new materials and medicines discovered with its help.