Meet the British startup on the brink of winning the race to produce the world's first commercial-grade quantum chips.
Forward: features are independent pieces written for Mewburn Ellis discussing and celebrating the best of innovation and exploration from the scientific and entrepreneurial worlds.
The dawn of quantum computing is here. Chips a trillion times more powerful than classical machines are entering prototype phase.
It is something of a surprise, therefore, to find a small British company rivalling the likes of Google and IBM in the race to build a truly scalable quantum chip. Quantum Motion is a UCL spin-out generating serious headlines.
Last year it announced a breakthrough, isolating a single electron and measuring its quantum state for nine seconds, described at the time as “an unimaginably long stretch”. Intel, for comparison, talks of times of one second.
There are two reasons for believing Quantum Motion may be the frontrunner to produce a quantum chip for mass production. The first is the calibre of the founders. John Morton is professor of nanoelectrics at UCL and author of more than 120 research papers on quantum technology. He's been a leader in the industry for two decades.
Quantum Motion CTO John Morton standing next to a dilution refrigerator.
His co-founder is Simon Benjamin, professor of quantum technologies at Oxford University. The pair are a perfect match. Morton is the experimental scientist, able to design the hardware. Benjamin is the expert on quantum theory.
The second reason is the approach the company is taking. Whereas rivals are focussed on esoteric ideas such as photon gates and trapping atoms of caesium in a laser grid, Quantum Motion is using traditional silicon chip processes.
“There are all kinds of approaches to building a quantum chip,” says Morton, speaking at the company HQ on the Caledonian Road in north London. “Some are based on exotic materials like superconducting qubits and majorana fermions. We are trying to make devices similar to silicon transistors, and use single electrons trapped within those silicon transistors as the qubit.”
Quantum Motion cools the chip down to -273C, just above absolute zero – a notable upgrade on a PC cooling fan. The quantum architecture is built with standard silicon hardware.
This means the cost of production will be slashed. Chip production is notoriously expensive. The new Texas Instruments semiconductor wafer plant in Sherman, Texas is budgeted to cost $30 billion. Intel's new plant in Ohio will cost $20 billion. When Quantum Motion enters production it can do so using plants such as these.
“If you really want to scale this technology you will need to scale to millions of qubits,” says Morton. “There are very few technologies that make millions of anything, with the exception of the silicon transistor. Our view is, if you can make qubits out of silicon, why would you do it any other way?”
Silicon chip manufactured at CEA Leti containing an array of devices, wired bonded onto a Quantum Motion board, used to trap single electrons and measure their quantum states
So how close is Quantum Motion to building a workable quantum chip?
“There are no major roadblocks,” reveals Morton. “Many people predict a fully error corrected quantum computer with a million qubits by the end of the decade. But you know what? We are going to see some interesting milestones along the way.”
To speed up development, Quantum Motion aims to build smaller quantum systems working in parallel.
“It is what we call multi-core quantum computing,” explains Morton. “Our 100 qubit processor will be about a few microns in size. This is about a million times smaller than typical superconducting qubits. At that stage we'll be ready to run many copies of 100 qubit machines on the same chip. That will allow us to solve problems in parallel. That's on a five-year horizon.”
Even 100 qubits pack a punch. The logic of quantum computing means that small numbers of qubits offer immense processing power. In a classical computer, a bit can hold a value of 1 or 0. Adding more bits equates to linear improvement. By contrast, a qubit can harness quantum properties to be both 1 and 0 at the same time. Thus each additional qubit doubles the processing power of the system – an exponential gain. At just 275 qubits it is possible to hold more computational states than there are atoms in the universe.
The multi-core model means Quantum Motion can create a powerful quantum computer before mastering a thousand or a million qubits on a single system. Commercial applications can be run at an earlier stage and via a single-core approach.
It also opens the door to mass participation in quantum computing.
“If you build a quantum data centre, you don't want a building that is just one quantum computer. You want many, many quantum computers. Otherwise it will be prohibitively expensive. If your data centre has thousands of quantum computers in racks, working in parallel, it is open to all kinds of different applications.”
The public will be able to access quantum computing via the cloud.
“It will be like Siri on the iPhone,” says Morton. “The decoding of your voice by Siri isn't done on the phone. It's sent to servers on the cloud and the computation is done there and the result sent back. As far as the user is concerned, they don't care whether the work is done on their phone or on a quantum data centre.”
The Quantum Motion data centre will serve enquiries, from companies to universities, and send the results back. Quantum as a service.
The applications of quantum chips
What will quantum computing be used for? The question is so broad Professor Morton pauses before answering.
He makes it clear that quantum computers are a radical break with any current machines.
“It's not right to compare them with classical computers. Quantum computers will be so much faster they will tackle problems for which we don't even try to use computers today.”
The optimal problems are those with simple premises, but vast mathematical variables. “There's a decades old materials modelling problem that has just 25 atoms, arranged in a 5x5 grid. The problem may sound simple with so few atoms, but solving it accurately is way beyond any classical computer today. But if you feed it into a quantum computer it is able to use its vast computational space to figure it out.”
It will be possible to model the interaction of atoms and molecules. In biology, a molecule called FeMo cofactor (FeMo-co), which bacteria deploy to extract nitrogen from air and create ammonia, is of profound interest. It suggests there is a more efficient way to generate environmentally friendly ammonia. A quantum computer of 10 million qubits could simulate FeMo-co in ten days to unlock its secrets. A classical super-computer is unable to decode the problem within any meaningful timeframe.
“We are already seeing the first applications modelling the world around us,” says Morton. “This means making better materials, catalysts and drugs. Quantum computers will design chemicals and materials in ways that it's just not possible to at the moment.”
Ultimately, the applications are so varied we cannot yet conceive of the range of future uses, says Morton: “Look at the 1960s and valve computing,” he says. “No one could have predicted rideshare and taxi apps running on smartphones. The best thing we can do is to open up the application space to millions of people and see what they build.”
The commercial potential
In November 2021, Quantum Motion entered a consortium to build a fully functional quantum computer control chip in silicon, running in deep cryogenic temperatures. Named Altnaharra, the project includes Oxford Instruments who supply cooling equipment, the UK National Physical Laboratory, Oxford Ionics – a start-up developing quantum computers using trapped ions, and scientists at the University of Glasgow with expertise in superconducting qubits. The key challenge is to develop the control and measurement electronics which are able to run at the same low temperatures as the qubits.
The project also highlights the intellectual property challenge for an innovator. Morton admits he's aware of the IP trade off between openness and protecting IP. “We publish as much as possible, but it's important for us to look at everything we do and protect it. An IP portfolio is an important way for us to show investors that we are generating value.”
He adds it's also vital to keep publishing in order to attract talent: “It's very challenging to hire as there is a limited number of candidates. Quantum companies will survive or fail based on the talent they bring in. It's by talking about what we are doing that we will encourage people to come and work with us.”
It's pleasing the ethos in the quantum world is collegiate. An academic philosophy unites competitors. “There's a lot of mutual respect between companies,” says Morton. “Even within big companies these will be people who have probably interacted with me and others here as academics, maybe ten years ago. So there's quite a lot of support across the sector as everyone is facing the same challenges.”
As a mass-production draws near, ferocious commercial pressures are likely to change the mood. With tens of billions of dollars in store for the company that dominates the quantum sector, amiable consensus is unlikely to work.
Morton won't be drawn on his financial aspirations – he's raised £20m so far in equity funding and grants, and will be embarking on another fund-raising round this year. Headcount will rise from the current 25 full-time staff. Unicorn status (a billion dollar valued start-up, in Silicon Valley parlance) is plausible.
Ultimately, Morton and Benjamin want to take Quantum Motion all the way to full chip production. “My co-founder and I are totally united behind this vision of building quantum computers.” Neither wants to sell to a tech major. “We value our independence a lot.”
Morton laughs at the idea he could be the Elon Musk of quantum chips. Perhaps a better parallel is Demis Hassabis, founder of another N1 postcode company – DeepMind. The neural networks of DeepMind also solve previously unapproachable problems. DeepMind's chess engine AlphaZero took 24 hours of self-training to defeat rival world-champion programs Stockfish and Elmo. In the field of protein-folding, mapping a single protein was the work of a PhD. DeepMind is set to decode 130 million proteins inside a year.
Professor Morton says that the two enterprises sit at opposite ends of the information spectrum. The neural networks built by DeepMind require immense data volumes to learn from: “They deal with very large amounts of information. Whereas quantum computers start with very small inputs, that have lots of different possibilities, and then a very simple output.”
Has he met Hassabis? “We haven't,” says Morton with a lament. “Although he was in UCL Computer Science. And DeepMind is just around the corner in Kings Cross.”
They should meet. DeepMind changed the AI industry. Quantum Motion looks set to change the chip industry.
And maybe Hassabis can prepare Morton and Benjamin for the global publicity that will follow as quantum computing starts to change science, finance, and life as we know it.
Even the modest Morton can see the attention he's going to generate if Quantum Motion beats the big tech companies to dominate quantum computing.
“We are building the most powerful computer the laws of physics allow,” he reflects. “There's not much more you can do to hype that.”
Quantum computers are now a reality
Andrew Fearnside, Senior Associate and Patent Attorney, Mewburn Ellis comments:
Quantum computers are no longer confined to theoretical musings and, although currently at prototype stage, they are now a reality. By harnessing the unique properties of quantum states, scientists and engineers have developed processors able to perform quantum computations. That said, the step from prototype to viable commercial product is a large one. It may be the step at which some quantum computing start-ups stumble. As a strategy for commercialising quantum technology at scale, the team at Quantum Motion can take pride in the foresight they have shown in tackling these challenges at the outset. Clearly, this is a company to watch.
Written by Charles Orton-Jones
Images supplied by Quantum Motion
