This summer Oxford Quantum Circuits became the latest quantum computing company to make their quantum hardware available to users over the cloud — the first UK company to offer quantum-as-a-service.
The Reading-based startup is joining the ranks of the big Silicon Valley tech companies Google, Amazon and IBM in offering QCaaS. While its system can’t quite compete with the number of qubits that these big challengers have to offer, it is a sign of how quantum computing is coming into the mainstream.
Let’s be clear — quantum computers aren’t terribly useful right now. Despite a few lab demonstrations of quantum supremacy — where quantum computers perform a function better than a classical supercomputer — in the real world companies are still struggling to find a compelling use case.
Companies like OQC are hoping that their quantum-as-a-service offerings will change that.
The pressure is on
Quantum computing companies are coming under pressure to prove their worth. Over the past year, we’ve seen a quantum frenzy.
Investment is at an all-time high: overall investing in the sector has reached $2.5bn so far this year, up from $1.5bn in all of 2020 and less than $500m each of the previous year, according to financial data firm PitchBook.
But much of the quantum computing capability remains severely underpowered. OQC’s quantum computer, which is based on superconducting qubits, cooled down a temperature close to absolute zero, has only 4 qubits, compared with the 65 that IBM’s Hummingbird can give you access to.
However, OCQ’s CEO Ilana Wisby argues that the high quality of the qubits at least in part compensates for that: "With this 3D proprietary qubit architecture invented by OQC's founder Peter Leek at the University of Oxford, called the Coaxmon, we have demonstrated incredible coherence and incredibly low cross-talk by taking all of the control wiring that usually networks the qubits on the same plane.”
Some remain sceptical about this. Olivier Ezratty, a French consultant and author of Comprendre l’informatique quantique (an English version is to be published in September) points out that OQC’s press release “didn’t come with a scientific paper on their hardware product nor did they give any more detail about the QCaaS offering”.
It is a tendency that Ezratty has seen a lot lately. “Since the beginning of the year, more and more startups are jumping on the quantum bandwagon yet they have in their hands a piece of hardware that is not ready to compete with what is on the market”.
Ilana Wisby evades this issue but instead elaborates on the encouraging results that OQC’s first client, Cambridge Quantum (CQ) has been able to show with its IronBridge cybersecurity platform. “We ran 5,000 programmes per second with a total of 10 million programmes execution”, she claims. “The results of the application will be published as part of a wider publication by Cambridge Quantum.”
Wisby also told Sifted that OQC was working on a next deployment with many more qubits, that they will soon make public.
A market expected to be worth $52bn by 2030
The market for quantum as a service is tiny at the moment less than $50m in 2020 and mainly comprised of research and consulting projects on one hand, and free experimental access to platforms on the other according to The Quantum Daily, the media site focused on quantum technologies.
However, interest from potential end users is growing. QCaaS revenue is expected to reach $4bn by 2025 and $26bn by 2030, according to a report from The Quantum Insider, the data offering of The Quantum Daily.
Moreover, these numbers don’t include the revenue generated by quantum computing application providers, which is expected to bring the total market size to more than $50bn, according to the report.
Amazon, Google and Microsoft vs the private cloud model
At present, QCaaS has been a US-focused game, dominated by AWS subsidiary Amazon Braket, Google AI Quantum and Microsoft Azure Quantum.
There are several different models.
Amazon Braket gives access to its own simulator as well as to other quantum computers, including D-Wave’s “annealer”, Rigetti’s “superconducting processor” and IonQ’s “trapped ion processor”, all on a pay as you go pricing model.
Google has adopted the same model for its quantum cloud platform, where companies can experiment their algorithms on Google’s own superconducting processor, IonQ’s trapped ion processor, and AQT’s trapped ion and Pasqal’s neutral atoms simulators.
Microsoft Azure Quantum, at the other end, works on an hourly or monthly basis for IonQ’s and Honeywell’s trapped ion processor and QCI’s superconducting processor.
The rest of the market is comprised with a handful of other quantum hardware companies and startups who have chosen to deploy their own cloud platform, IBM Q being the first to come and the most advanced.
With the arrival of European quantum-as-a-service companies like OQC, and quantum computers being set up in countries like Germany and Finland, there should be increasing opportunities for companies to access quantum machines.
With more experimentation, perhaps, soon, someone will discover good business use cases for the technology.