Quantum computing is a way of processing information with qubits, which follow the rules of quantum physics instead of the simple on-off logic of ordinary chips. For a narrow set of problems, a large enough quantum computer could finish work that no regular computer can handle. Examples include simulating molecules and breaking certain kinds of encryption. As of October 2026, though, today’s machines are still small, error-prone and mostly used for research.
Here is what that means in plain English, what the machines can and cannot do, and how close they are to being useful. For related explainers, browse our science and technology guides.
Bits vs. qubits
Your laptop stores everything as bits. Each bit is either 0 or 1, like a light switch that is off or on. Billions of these switches, flipped very fast, run every app you use.
A qubit is different. Before it is measured, a qubit can be in a blend of 0 and 1 at the same time, a state physicists call superposition. When you measure it, the blend collapses into a plain 0 or 1, with odds set by the calculation.
Qubits can also be linked through entanglement. Then the state of one qubit is tied to the state of another, so the group behaves as a single system. Because of this, adding qubits grows the space a quantum computer works in exponentially. Ten qubits describe 1,024 possible states at once, and 300 qubits describe more states than there are atoms in the observable universe.
That does not mean a quantum computer simply tries every answer in parallel. Instead, a quantum algorithm uses interference, much like waves that cancel or reinforce each other. Wrong answers cancel out, right answers grow stronger, and a measurement then picks out a likely solution.
How a quantum computer is built
There is no single design yet. Several kinds of hardware compete:
- Superconducting circuits. Tiny loops of metal cooled to near absolute zero. Google and IBM use this approach.
- Trapped ions. Charged atoms held in place by electric fields and controlled with lasers.
- Neutral atoms. Uncharged atoms arranged in grids by beams of light.
- Photons. Particles of light routed through optical chips.
The superconducting approach traces back to experiments from the mid-1980s. In 2025, John Clarke, Michel Devoret and John Martinis won the Nobel Prize in Physics for that work. They showed quantum tunneling and distinct energy levels in an electric circuit large enough to hold in your hand. That work underpins the qubits in many of today’s machines.
Whatever the hardware, qubits are fragile. Heat, vibration and stray signals knock them out of their quantum state, which causes errors. So most machines sit inside large refrigerators or vacuum chambers, and much of the engineering goes into keeping noise out.
What running a quantum program looks like
A quantum calculation follows a few steps. First, the machine sets every qubit to a starting value. Then it applies a sequence of operations, called gates, using precise pulses of microwaves or laser light. Finally, it measures the qubits and reads out a string of 0s and 1s.
Because the result is probabilistic, one run is rarely enough. The program repeats thousands of times, and the most frequent answers point to the solution. Meanwhile, an ordinary computer handles the setup, sends instructions and checks the output. In practice, quantum chips work as accelerators attached to classical computers, not as standalone replacements.
Error correction is the real race
Because single qubits fail so often, researchers combine many physical qubits into one more reliable “logical” qubit. The idea is similar to repeating a message several times so the listener can spot a mistake. Then the system can detect and fix errors as it runs.
For years the catch was that adding more qubits added more errors. In December 2024, Google reported that its 105-qubit Willow chip cut the error rate roughly in half each time it enlarged its error-correcting grid. The company called this the first system to operate “below threshold,” meaning bigger codes made results better rather than worse. On a benchmark task, Willow finished in under five minutes a computation Google estimated would take a top supercomputer 10 septillion years. That benchmark has no practical use, however.
IBM has set a public goal. In June 2025 it announced plans to deliver a fault-tolerant machine called IBM Quantum Starling by 2029. It is designed to run 100 million operations on 200 logical qubits, at a new data center in Poughkeepsie, New York. Other companies have published their own roadmaps, and their dates differ.
What quantum computers could do
Quantum computers will not replace laptops or phones. They are specialized tools, and the strongest use cases share one feature: the problem itself follows quantum rules or has a hidden mathematical structure.
| Area | What a quantum computer might help with | Status as of October 2026 |
|---|---|---|
| Chemistry and materials | Simulating molecules for batteries, fertilizers and drugs | Early experiments; useful scale not yet reached |
| Physics research | Modeling magnetic materials and exotic states of matter | Most active area of real results |
| Optimization | Routing, scheduling and portfolio problems | Advantage over classical methods still unproven |
| Cryptography | Breaking RSA and elliptic-curve encryption with Shor’s algorithm | Requires far larger error-corrected machines than exist |
| Machine learning | Faster training or sampling for some models | Largely theoretical |
Artificial intelligence, by contrast, runs on ordinary chips today and does not need quantum hardware. If the difference between AI terms is unclear, our explainer on artificial intelligence vs. machine learning sorts them out.
Why it matters for encryption now
Much of today’s online security, from bank logins to software updates, relies on math problems that ordinary computers cannot solve in any reasonable time. A large fault-tolerant quantum computer could solve some of them. That machine does not exist yet. Still, attackers can record encrypted data today and try to decrypt it later.
So governments are moving early. The National Institute of Standards and Technology (NIST) acted first. In August 2024, it published its first three post-quantum encryption standards, FIPS 203, 204 and 205, and urged organizations to start switching. On its post-quantum cryptography project page, NIST says it plans to deprecate and then remove quantum-vulnerable algorithms from its standards by 2035, with high-risk systems moving sooner.
For most people, this change will happen quietly through browser, phone and operating system updates. Keeping devices updated is the practical step.
Is quantum computing the next big thing?
It might be, but on a longer timeline than headlines suggest. The science has moved fast since 2019, when researchers first showed a quantum chip beating a supercomputer on a contrived task. Error correction is now working in the lab, and companies have published dated roadmaps.
That said, several hard problems remain:
- Useful machines may need hundreds to thousands of logical qubits, each built from many physical ones.
- Cooling, wiring and control electronics must scale along with the chips.
- For many tasks, nobody has yet proven that a quantum computer will beat the best classical method.
In practice, expect the first clear wins in chemistry and physics research, not in consumer products. Most people will reach quantum computers, if at all, through cloud services rather than owning one. Bold claims that quantum computing will soon transform everything deserve some caution. Real progress is steady, but it is measured in years, not months.
Frequently asked questions
What is quantum computing in simple terms?
It is computing with qubits, which can hold a blend of 0 and 1 and can be linked together. For certain problems, this lets a quantum computer explore possibilities in ways ordinary computers cannot.
Is quantum computing the next big thing?
Possibly, for specific fields like chemistry, materials and cryptography. As of October 2026, though, machines are still experimental, and large error-corrected systems are targeted for the end of the decade.
Can I buy a quantum computer?
Not in a practical sense. The machines need special cooling or vacuum systems. Researchers and businesses usually rent time on them through cloud services offered by the companies that build them.
Will quantum computers break passwords and encryption?
A large fault-tolerant machine could break some common public-key encryption, but none exists yet. NIST has already published replacement standards, and software makers are rolling them out.
Have a question about this topic? Ask in the comments below. Our editors read every comment and answer questions here, and we add useful answers to the guide.