The cloud is big business – and it still has much growth potential. The global market for public cloud services this year will be worth about $597bn (£462bn), according to an estimate by Gartner in April. The tech research giant expects the total to leap to $725bn in 2024 and believes that this rate of expansion is sustainable at least in the next few years. It has forecast that three-quarters of organisations “will adopt a digital transformation model predicated on cloud as the fundamental underlying platform” by 2026.
But what if the massive processing power of the cloud providers’ servers could be easily outstripped by that offered by an alternative technology? Quantum computing can reportedly perform some functions in microseconds that, as things stand, would take several millennia for a conventional computer to wrestle down. Does that so-called quantum advantage mean that demand for classical computing could decline dramatically over the next few years?
We need to set the right expectations about quantum computing. It will be able to solve only certain types of problems more efficiently
“Computing has become increasingly homogeneous over the past two decades,” notes Richard Hopkins, a distinguished engineer specialising in hybrid cloud solutions at IBM and a former senior quantum ambassador at the company. “We’re going to see more speciation, because we need it. In essence, there are lots of very powerful PCs underneath the cloud. We can’t afford to try solving large-scale optimisation problems using conventional computers such as these, because we’d be consuming huge amounts of power.”
Bloomberg has estimated that cloud could be using 8% of all electricity generated in the world by end of this decade. One of the reasons for this, according to the Institute of Electrical and Electronics Engineers, is that nearly all of the possible efficiencies have been wrung out of classical computing.
Hopkins says that IBM expects quantum computing to be solving complex problems, without requiring anywhere near the same amount of power, before 2030.
How quantum computing is progressing
Quantum computing still has some significant development hurdles to surmount. For instance, the computers require cooling to -273ºC – absolute zero – and have other demanding operational requirements. Despite this, some quantum capability is already available. What’s more, it’s accessible via the cloud.
“We have launched Amazon Braket, a fully managed service giving customers access to different types of quantum hardware,” says Simone Severini, director of quantum technologies at Amazon Web Services and professor of physics of information at University College London.
He reveals that the service is still limited in scope, but “the idea is to allow experimentation and to help speed up scientific research and software development for quantum computing”.
Hopkins believes that quantum computing “won’t enter the mainstream for some years, but today at IBM we have 24 quantum computing systems in our cloud. Our System Two, which can run quantum processing units alongside classical central processing units, is set to be released later this year.”
How quantum computing could benefit businesses
But what are quantum computers able to do? Hopkins explains that quantum computing applies to problems beyond the capability of the algorithms that can be run on classical computers. “For example, in credit card processing, we used a quantum algorithm alongside a classical one and were able to reduce the number of false positives and negatives beyond the state of the art.”
Severini says that “some of the hardware available on Amazon Braket is universal – that is, it tackles any computing problem – and some of it is specific to particular problems”.
Computing has become homogeneous over the past two decades, but we’re going to see more speciation – because we need it
He stresses that quantum computing should not be seen as a panacea. There are some problems that are extremely hard for today’s conventional systems to solve that will also be extremely hard for quantum computers, such as challenges in combinatorial optimisation. This is the term applied to tasks such as solving the so-called travelling salesman problem, which involves calculating the best route for a salesman to visit all the cities on his schedule once and then return home. As the number of destinations increases, working out the most efficient itinerary gets very computationally intense. Combinatorial optimisation has numerous commercial applications, such as finding the best way to cut up sheets of raw material in a factory so that waste is minimised.
“We need to set the right expectations about quantum computing,” Severini says. “It will be able to solve only certain types of problems more efficiently. Uploading data to quantum computers can be very time-consuming and the notion of quantum storage is still unclear. It’s potentially useful when the input is small and the number of computing steps required is very large.”
He doesn’t expect quantum computing to become primarily a business service. In his view, it “will be used specifically as a research instrument and have the most impact on areas to do with quantum physics itself”.
Research, though, could cover areas such as molecule interaction simulation, pricing optimisation and supply chain optimisation, all of which have commercial potential.
Does quantum computing pose a threat to data security?
There are security concerns about the potential misuse of quantum computing. One example of ‘small input, many steps’ is working out the prime numbers that have been used in a cryptography key. With a 400-digit key, a classical computer running a million factorisations per second could, over the lifetime of the universe, try out 1024 possibilities. To be sure of cracking the key, it would need to examine 10200 potential combinations. A quantum computer, by contrast, could feasibly break the long-standing RSA cryptography standard on which much of today’s online security systems still rely.
The authorities are taking this risk seriously. For instance, the US government’s National Institute of Standards and Technology announced the first four quantum-resistant cryptographic algorithms in 2022 after a six-year competition.
Investors are also seeking solutions to the problem. Kamil Mieczakowski, a partner at venture capital firm Notion Capital, reports that it has invested in Arqit, “a company that helps enterprises to become ‘quantum ready’ with its quantum-proof encryption technology”.
Mieczakowski says that it’s still early days for quantum computing and adds that, once it does become more widely available, it’s likely to be “overkill” for most purposes. Classical cloud computing will still be the main workhorse.
He observes that “it’s very expensive to build and run quantum computers. They’re inherently fragile too. The cloud will be the cheapest way to access quantum computing capability.”
Quantum computers will, in effect, become a part of the cloud as it continues to grow and add new tech that can process data more quickly and energy-efficiently. But quantum computing is unlikely to usurp the cloud as the “fundamental underlying platform” on which tomorrow’s businesses will be built.
The cloud is big business – and it still has much growth potential. The global market for public cloud services this year will be worth about $597bn (£462bn), according to an estimate by Gartner in April. The tech research giant expects the total to leap to $725bn in 2024 and believes that this rate of expansion is sustainable at least in the next few years. It has forecast that three-quarters of organisations “will adopt a digital transformation model predicated on cloud as the fundamental underlying platform” by 2026.
But what if the massive processing power of the cloud providers’ servers could be easily outstripped by that offered by an alternative technology? Quantum computing can reportedly perform some functions in microseconds that, as things stand, would take several millennia for a conventional computer to wrestle down. Does that so-called quantum advantage mean that demand for classical computing could decline dramatically over the next few years?
We need to set the right expectations about quantum computing. It will be able to solve only certain types of problems more efficiently
