Quantum Untangled: The ‘as-a-service’ future for quantum computing
Quantum Untangled: The ‘as-a-service’ future for quantum computing
The hyperscalers surfing the quantum wave on startup surfboards.
The general wisdom in the burgeoning world of quantum computing is that the cloud is king. Quantum computers are, after all, incredibly complex and expensive machines to run, let alone build. Making the world pay a premium on accessing the dozen or so machines that do exist, therefore, makes more sense than striving to manufacture a fleet of these exotic computational engines for every single company on the planet confronted with intractable mathematical problems.
Most of the big players in quantum computing agree. IBM is a leader in this area, heavily marketing its own quantum cloud concept and providing access to this new form of networking to a multitude of aspirant students and entrepreneurs. All of the hyperscalers have some form of quantum product, too, powered either by their own devices or those rented from the burgeoning ecosystem of quantum startups.
This latter form of ‘quantum-as-a-service’ is a definitional fudge, argues Richard Murray, CEO of British photonics firm ORCA. Ideally, he told me during a recent interview, hyperscalers like AWS, Azure and Google Cloud should be willing to spend the money to build their own quantum computers and run them from their own data centres, instead of relying on plucky startups to shoulder the hardware burden. “To us, it is not representative of true cloud if the quantum system is not within their data centre,” Murray says, adding that ORCA had been approached more than once to run its own quantum hardware as an ancillary feature to a hyperscaler’s broader cloud offering.
Originally formed out of a quantum sensor lab at the University of Oxford, ORCA builds room-temperature quantum computers that are proving useful when paired with GPUs to train image generation AI models. The first version of the P-1 was sold to the UK’s Ministry of Defence and, since then, they’ve sold another four machines. Work is also underway to speed up production time and work on adding more photons to the chips.
“They are still highly specialist devices,” says Murray, hardware that requires theoretical physicists to be on site to help the engineers assemble all the different pipes, chips and sprockets together into something resembling humanity’s most advanced computational engine. Suffice to say, none of that has been easy. “We're trying to enter into a regular product delivery mode,” says Murray. “But,” he adds, a little baldly, “we are still working with a quantum computer.”
While most quantum computing is delivered through the cloud, there are other methods companies are using to get the power of a qubit into the hands of end-users. For some organisations, having the machine on location, where they can safely syphon secure data through the machine, is vital. Others, like Oxford Quantum Circuits, have sold machines directly to data centre providers like Equinix. For its part, IBM has placed its System One quantum computer in specific facilities like the Cleveland Clinic for medical research. Meanwhile, Universal Quantum, based out of the University of Sussex, has also been commissioned by Germany's Aerospace Centre to build a pair of quantum computers as part of a €67m contract.
Does it matter that some hyperscalers piggyback off the quantum computers run by startups? Yes and no. Providing access to quantum hardware within the native cloud infrastructure can allow for more rapid scaling to meet growing demand, and for regionalisation to cover regulatory bases. It also allows for improved latency and security, removing the risk of sending data off to a third-party network.
But the problem is that, right now, there are more types of quantum computer out there than use cases. From a hyperscaler’s point of view, buying in hardware from IonQ, ORCA, Quantinuum or any of a multitude of startups would inevitably incur significant outlay and risk. Right now, quantum computers are expensive to maintain and often requiring supercooling, and, for the most part, are currently little more than immature research machines waiting for a major breakthrough. If the demand was there from end-users, the hyperscalers would be knocking down ORCA's door to buy a fleet of P-1s to install. Right now, that demand isn’t there. Until that point, there are plenty of quantum startups happy to provide the hyperscalers with access to their hardware and earn revenue while end-users work out just what to do with this quantum power.
Partner content
The security challenges of digitalising the energy grid - Tech Monitor
How do we restore trust in the public sector? - The New Statesman
Brands must seek digital fashion solutions - Tech Monitor
The new to direct capital deployment to decarbonise household electrification - Capital Monitor
What factors drive data centre vendor selection? - Tech Monitor