Quantum computers may no longer be the stuff of science fiction, but how far is the UK from fully realising their world-changing potential?
In 2013, the UK government set aside 拢270 million to invest in quantum technologies, aiming to help produce the world鈥檚 most powerful working quantum computer. While it couldn鈥檛 compete with the billions of dollars that tech giants like IBM and Google have poured into quantum computing research, the approach still seems to have paid dividends.
鈥淭he UK was pioneering in terms of its quantum strategy,鈥� said , CEO of Oxford Quantum Circuits (OQC), at the National Quantum Technologies Showcase (NQTS) in London on 11 November. 鈥淭hat meant that there was a ton of really early innovation and leadership in the quantum space from the UK at the academic level and that鈥檚 fed through, so we now have a really vibrant quantum ecosystem.鈥�
Advertisement
In September, OQC, which operates the country鈥檚 only commercial quantum computer, announced that it had teamed up with data services company Cyxtera to let , starting next year. The computer鈥檚 current chip, which runs on superconducting quantum bits (or qubits) similar to those used by Google and IBM, has eight qubits.
This small number of qubits 鈥� and this also applies to the hundreds used by IBM鈥檚 Osprey, the world鈥檚 most powerful quantum computer 鈥� means that quantum computers can鈥檛 yet run algorithms that are of direct, practical use.
That doesn鈥檛 mean nothing can be done on them, though, says Wisby. She says it is important to have the facilities available so people can learn how to use quantum computers ready for when the technology does become more powerful. 鈥淲hat people are starting to do is to break down big problems into smaller problems and then start to test them.鈥�
We are now in what has been named the noisy intermediate-scale quantum era, in which we are gradually increasing the number of qubits in machines, but they can鈥檛 yet reliably cope with small faults or errors that may creep in because of the difficult nature of making measurements of tiny quantum systems.
Quantum computers won鈥檛 start to be truly useful until they reach scales of about 1 million qubits, says at quantum software company Riverlane in Cambridge, UK. The current timescale for a machine of that size is around 10 years, going by road maps from companies like IBM, but things could happen faster if we find efficient quantum algorithms, she says.
The ability to correct errors will be crucial. Normal computers spot and fix their own errors, but this is a trickier problem in the quantum world because of how heat can disrupt cooled quantum systems and how the qubits, and thus their errors, are interlinked, through the strange phenomenon of quantum entanglement. This means errors can propagate through the system. Researchers and companies like Riverlane are working on technological solutions for error correction in preparation for when machines with lots of qubits do start getting made.
Large, error-corrected quantum computers are often referred to as digital quantum computers. Although they leverage the weird world of quantum mechanics, they still ultimately deal in 1s and 0s and you can run algorithms on them that are similar to those you would use on a conventional computer.
Some quantum computers rely on an analogue approach, instead, which relies on qubits being easily manipulated to create mini versions of real-world scenarios.
One UK example is from Glasgow-based laser company M Squared, which announced its neutral atom quantum computer, Maxwell, at NQTS. Neutral atom machines use a grid of lasers to keep atoms suspended in space that act as the qubits, in contrast to the superconducting metals used in qubits in IBM and OQC鈥檚 machines.
This means that, unlike superconducting quantum computers, which have their qubits in a fixed location, neutral atom machines can move them around and make them connect with other qubits that aren鈥檛 immediate neighbours. This interconnectedness allows the computers to run unique analogue algorithms.
鈥淭he hardware that we鈥檝e got on Maxwell is compatible with this emerging form of analogue and simulation-based systems,鈥� says , CEO of M Squared. 鈥淢any of the kinds of industrial problems that we have in the world are about industrial optimisation and simulation, and we see that the next stage is going to be based on the analogue algorithms and circuits that we built with Maxwell.鈥�
While these machines still require a certain amount of error correction to be functional, it isn鈥檛 at the same threshold as the digital machines, says Malcolm, so people might be able to get useful results out of them much sooner than with the large superconducting systems.
Multiple quantum computing approaches potentially have a role to play given the myriad applications that could benefit from them, says , chief scientist at the UK鈥檚 National Quantum Computing Centre. 鈥淎ll of them are part of the ecosystem, while we still have years to come to have a really scalable, useful, fault-tolerant quantum computer.鈥�
