What is quantum computing?

Quantum computers are often touted as the next generation of computing. They rely on the laws of quantum mechanics — the weird behavior of particles at a subatomic scale — to process information. Currently, quantum computers are too small, too difficult to maintain and too error-prone to compete with today's best classical computers. However, many experts expect quantum computing to one day overtake classical computing for specific tasks. 

The technologies that enable quantum computing have advanced rapidly in the past few years. One day, they may be able to solve problems that are too complex for even today's most powerful conventional computers. This massive performance gain could open the door to many exciting uses including in pharmaceuticals, climate modeling and manufacturing, all of which rely on hugely complex simulations.

Quantum computing vs classical computing

Classical computers process data using binary bits, which can be in one of two states — 0 or 1. The bits are encoded on transistors, which can be made from silicon, germanium or other semiconductors.

Quantum computers use particles such as electrons or photons that behave as quantum bits, or qubits, which represent a superposition of both 0 and 1 — meaning they can exist in multiple states at once. Qubits may also be encoded in semiconducting materials such as silicon, or even superconducting materials, such as spinel (MgAl2O4) and lanthanum aluminate (LaAlO3).

To fully realize quantum supremacy, quantum computers need different algorithms that take advantage of the unique way qubits encode and process data. Scientists are developing quantum algorithms, which have lower computational complexity, meaning they require less runtime or number of operations against conventional algorithms. However, quantum algorithms would need to be run on large, fault-tolerant quantum computers, which are not yet available.

Qubits: Qubits are quantum particles that are equivalent to binary bits in classical computers. Given that qubits can be in more than one state, they offer exponentially greater processing capabilities than binary bits if they can be stitched together and used to run calculations.

Qubits process data using quantum gates, which are analogous to binary gates in classical computers. However, unlike binary gates, quantum gates are reversible. Some binary gates lose data as information is processed through them, but quantum gates preserve it. Combined together, quantum gates form quantum circuits.

Superposition: The primary difference between qubits and binary bits is that the qubits operate in superposition, meaning the qubit can represent both 1 and 0 simultaneously. This superposition enables quantum computers to perform calculations in parallel by processing all states of a qubit at the same time.

Entanglement: Quantum entanglement is a phenomenon associated with superposition, in which two subatomic particles — or qubits in a quantum computer — are linked over space and time. They are physically separate but share information and interact simultaneously. Regardless of the distance between the particles, if one is observed, then the state of the other is known.

How powerful are quantum computers?

Quantum superposition and entanglement make the processing potential of a quantum computer so much higher than a classical computer.

Whereas adding more classical bits linearly increases how many calculations a computer can do, adding more qubits to a quantum computer exponentially increases its computing power — far outstripping a classical binary computer once there are enough qubits. Scientists estimate that a quantum computer with roughly 20 million qubits will achieve quantum supremacy —  the point at which a quantum computer solves a problem a classical computer cannot.

Related: Scientists just built a massive 1,000-qubit quantum chip, but why are they more excited about one 10 times smaller? 

However, quantum computers are still very experimental. For one, the superposition that creates qubits, and the entanglement that stitches them together, are very easily destroyed — because the qubits interact with the external environment and become entangled with it . When that occurs, the information they carry is lost or corrupted. That makes quantum computers extremely error-prone. To get around this, companies are deploying multiple approaches, such as supercooling to just above absolute zero and using electromagnets to isolate the qubits. 

How do quantum computers work?

Quantum computers have an iconic chandelier-like architecture, comprising a series of interconnected tubes and wires that host different layers of the computer. Most quantum computers are linked with massive, powerful refrigerators so the processors can be cooled to near absolute zero to mitigate thermal noise and vibrations. So many of the chandelier's layers work to get the quantum processor, housed near the bottom layer, very cold.

Quantum computers all have slightly different architecture, but they tend to have the following elements.

Quantum data plane: The quantum data plane houses the qubits and is where data is processed via quantum gates. The structure that holds the qubits in place differs between different types of quantum computers. Some qubits are made of solid superconductors cooled to just above absolute zero. Others use electromagnetic fields to trap ions, or charged atoms that act as the qubits, in high-vacuum chambers. The vacuum pressure minimizes interference from vibrations and stabilizes the qubits.

Control and measurement plane: The control and measure plane converts a digital signal from a classical computer into the analog signals used to change the states of the qubits in the quantum data plane.

Just as with the quantum data plane, quantum computers send signals in multiple ways, such as with microwaves or lasers.

Control processor plane and host processor: The control processor plane and host processor implement the quantum algorithm, which is a sequence of operations designed to run on a quantum computer to process the data. After performing a quantum calculation, the host processor ultimately provides a classical digital signal to the control and measurement plane.

Quantum software: Getting the processor output into the control and measurement plane requires another element: quantum software. Quantum computers require specially designed algorithms, which are most commonly described by a quantum circuit, or a routine that defines a series of quantum operations on the qubits. Quantum software is made up of quantum algorithms. Other quantum software is used to correct the errors generated when performing calculations on the qubits. 

Why do we need quantum computers?

In theory, quantum computers can potentially be far faster than classical computers and can simultaneously solve multiple complex problems. They are particularly promising for optimization tasks. Classical computers struggle or fail when a problem has an extremely large number of possible solutions. A quantum computer, however, could consider all potential solutions and quickly find the optimal one. Drug discovery or material science — where the fastest classical computers are currently deployed — are two examples of how quantum computers could be used.

Quantum computers could also transform artificial intelligence (AI). AI systems are trained using large data sets, so quantum computers could enable bigger and more complex data sets to be used for training AI, thereby leading to increasingly sophisticated systems.

Why are quantum computers so hard to build?

Quantum computers are delicate and susceptible to interference from external sources, such as temperature changes or stray particles. When there is interference, qubits are susceptible to decoherence — which is the collapse of quantum state. This decoherence makes quantum computers far more error-prone than conventional computers. While roughly 1 in 1 billion billion bits fail, the failure rate is roughly 1 in 1,000 for qubits.

Although there are ways to protect a quantum system from external influences, errors can still creep in. Even a single error can cause the validity of an entire computation to collapse. And because qubits are fundamentally different from bits, conventional error-correcting methods don't work.

Scientists have made quantum algorithms to compensate for errors in quantum computers, but these require qubits to run, reducing how many are available to process the data. Another quirk of quantum mechanics is that directly observing or measuring the state of a particle or atom in superposition destroys it. That means researchers must use tricky workarounds to read the quantum state of the output, as direct observation risks corrupting the data.

What are the implications of quantum computing?

Quantum computers will be a disruptive technology once we achieve quantum supremacy. But it's uncertain when scientists will build a quantum computer powerful enough — with millions of error-corrected qubits — and so far the most powerful quantum computers only have approximately 1,000 qubits.

Even then, classical computers will remain the easiest way to tackle most problems because they do not need to maintain quantum states. Quantum computers will likely only be used to tackle problems that are beyond the capabilities of classical computers.

One area that will likely be affected, however, is encryption, which protects sensitive data such as financial records and personal information. Modern encryption methods rely on mathematical problems that are too complex for classical computers to solve. However, a quantum computer’s processing power would easily be able to solve them. Quantum cryptography is now a burgeoning field, as researchers attempt to develop quantum-resistant encryption to protect sensitive data from being cracked by quantum computers in the future.