In a world driven by a quantum hype, many banks are now investigating whether the promised increased processing power of quantum computing could address problems that currently are too complicated to be solved. Liz Lumley investigates.
Tales of innovation are often peppered with stories of companies and industries that failed to see the impact new technologies would have on business models and customer behaviour. The evolution of digital environments has left many firms, financial services firms notwithstanding, struggling to play catch-up. Because of this, today’s technology conversations are more likely to be filled with the hypothesised promise of concepts such as Web3, cryptographically trusted currencies and the metaverse than with the day-to-day business of digital transformation.
But any new technology runs through a life cycle of theory, research, experiments and eventual placement in a production environment. Over the past few years, the hype around the eventual use of quantum-level computing within financial services has tended to go one of three ways: corporate venture capital arms investing in quantum computing tech companies; high-level theoretical mathematical research into preparing for a future where most cryptographically protected data and transactions are left vulnerable to the superior processing power of a quantum computer; and investigating business areas within the bank that would benefit from that increased power.
The increase in computational power from quantum is the result of the use of ‘qubits’ — the quantum version of bits used in classical computing. Quantum computers rely on logic governed by the unpredictable behaviour of subatomic particles, or ‘superposition’, in that its matter can theoretically exist in multiple states at the same time until it is measured.
This is radically different from the straightforward, binary logic of classical computing and the reason why some fear a future where quantum processing leaves most classical cryptographically protected systems vulnerable. However, that research lives mostly within the environs of academia and mathematical theory.
Separate from that research, many banks are now investigating whether the promised increased processing power of quantum computing could address problems that currently are too complicated to be solved.
Financial crisis
The industry is still recovering from the failure of risk models that lay at the heart of the 2008 financial crisis, where classical computing methods resulted in inaccurate business predictions. Today, many financial services firms process a high number of complex variables, such as portfolio optimisation and risk estimation, as well as financial market predictions that would benefit from increased computing power.
There is a big risk that quantum computing could break encryption protocols
Carlos Kuchkovsky, co-founder at quantum tech and Web3 start-up QCentroid, who spent 13 years at BBVA, works with several banks as well as fintech companies on preparing and developing quantum computing. A big part of what banks are investing time into is security and communication, he says. “There is a big risk that quantum computing could break encryption protocols,” says Mr Kuchkovsky. One of the top priorities for banks is to be ready for post-quantum cryptography in the next five years, he adds.
Business applications
The idea of the existence of a true quantum computer depends on whether you are speaking to a tech company like IBM, a bank investigating possible business applications for quantum, or a mathematician investigating a possible quantum future.
According to David Jao, professor in the department of combinatorics and optimisation at the University of Waterloo in Canada, a true quantum computer “does not exist”. Mr Jao’s university receives funding from the Royal Bank of Canada’s cyber security lab for research into post-quantum cryptography, focusing on a unique blend of pure mathematics and computer science that produces a data encryption so strong that quantum computers cannot crack it.
As a mathematician, Mr Jao is talking about a ‘cryptographically relevant quantum computer’, which refers to “a quantum computer that can usefully accomplish tasks that cannot currently be accomplished in the context of breaking cryptography”. A cryptographically relevant quantum computer “requires millions of physical qubits in order to implement the 5000 error-corrected logical qubits needed to perform cryptanalytic operations,” he adds.
IBM claims it has built more than 20 quantum computers. According to an IBM spokesperson, it put a quantum computer in the cloud in 2016. The IBM Quantum Experience — now the IBM Quantum Composer and IBM Quantum Lab — has an active community of more than 380,000 users, who have run more than one trillion circuit executions (more than four billion per day), on real hardware and simulators. This has led to the publishing of more than 1500 third-party research papers, according to the spokesperson.
By 2023, IBM will debut the 1121-qubit IBM Quantum Condor processor, incorporating the lessons learned from previous processors while continuing to lower the critical two-qubit errors so that they can run longer quantum circuits. The spokesperson says: “We think of Condor as an inflection point, a milestone that marks our ability to implement error correction and scale up our devices, while simultaneously complex enough to explore potential quantum advantages — problems that we can solve more efficiently on a quantum computer than on the world’s best supercomputers.”
Developing field
Banks are also running portfolio optimisation and simulations, looking at credit scoring and asset allocation in order to manage capital adequacy regulations, says Mr Kuchkovsky. A third area banks are investigating involve Monte Carlo simulations, which help to explain the impact of risk and uncertainty in prediction and forecasting models. He says a lot of companies are running quantum-level Monte Carlo experiments “because quantum random numbers are the only true random numbers for Monte Carlo”.
A lot of these experiments are an extension of research into machine learning. With quantum, “all the things that banks are using machine learning for, you can optimise three times faster with less energy and with better accuracy,” he says.
One of the challenges for banks moving their “code labs” to production is connecting productivity systems to quantum hardware through the cloud, says Mr Kuchkovsky. “Moving things from proof of concept to production, you need to pass through a lot of controls of banks, not only internal controls but also regulation.” This process can take almost three years, he adds.
QCentroid is now working with different quantum providers, start-up research firms, fintech companies and a marketplace of quantum solutions mostly in the field of environmental, social and governance and sustainability, according to Mr Kuchkovsky. This ecosystem can “adopt quantum computing faster than traditional companies, because they have the ability to connect in a faster way, as they don’t need to deal with all the risk and legacy challenges applied to a corporation,” he adds.
Many of the fintech companies Mr Kuchkovsky’s company is working with are in the cryptocurrency space. “They have the same goals as a bank: they need to try to make as much profit as possible with the lowest rates possible,” he says.
Nascent technology
Chintan Mehta, chief information officer and head of digital technology and innovation at Wells Fargo, also believes quantum technology is in its “early days”.
“You could liken it to where classical computing was in the 1950s,” he adds. “So, yes, it’s still pre-industrial. But we see enormous potential in the technology and believe it is essential that we, as the financial services industry, prepare for a post-quantum world.”
We see enormous potential in the technology and believe it is essential that we, as the financial services industry, prepare for a post-quantum world
As part of the bank’s artificial intelligence (AI) academic research programme, Wells Fargo started investing in the potential of quantum technology in 2019.
“We established a research partnership with the MIT–IBM Watson AI Lab in 2019, building on our academic relationships across our work in AI and other areas,” says Mr Mehta. “This partnership afforded us the opportunity to explore quantum computing in an advanced capacity using real-world problem sets.”
Currently, Wells Fargo is exploring and expanding basic building blocks to test drive mathematical computations in a quantum environment. These include exploration into approaches that lead up to vector mathematics and generalised linear algebra to test large-scale problem solving — for example, testing the performance of quantum computing against classical computing to support rapid recalculation pricing for a large book of trades.
Managing cyber security risk will also be at the forefront of the applications for quantum computing in the near term — ensuring the industry is prepared against potential risks in post-quantum cryptography. Wells Fargo is also experimenting with applications related to high-frequency trading, training neural networks.
“Much of our work is focused on testing speed and accuracy of parallel, or quantum, computing versus classical computing,” says Mr Mehta. “We see real potential for rapid scaling of large, complex data sets — exponentially reducing the time taken versus today’s techniques. But at the same time, we are actively testing for accuracy as the technology develops, given the significant physics challenges involved in realising the potential of quantum as a platform.”
Industry first
CaixaBank claims it was the first bank in Spain to work with quantum computing “and one of the first in the world to incorporate quantum computing into their R&D activity,” says Xavier Rebes, director of innovation at CaixaBank.
“Regarding the maturity level of quantum technology, it’s not quite at the mature point where any data scientist could use this kind of technology in their pipeline, but the goal is clear for most technology vendors,” he says.
Despite this, the bank is “very passionate about this technology”, Mr Rebes notes. In 2018, CaixaBank set up a team of experts with IT technicians, mathematicians and risk analysts dedicated to innovation in the quantum field, to explore the potential for quantum technology in areas such as: risk assessment and tail risk simulators; fraud detection with AI and machine learning; quantum-safe cryptography; portfolio selection and allocation; and data mining optimisation.
“Certain optimisation problems are solved using heuristics ([rule-of-thumb] shortcuts that enable problem solving that can be prone to errors); quantum computing enables us to locate the best solution without heuristics,” says Mr Rebes. “[It] also helps us find the best solutions in very short times, giving a competitive edge to the institutions that use this technology.”
CaixaBank’s first proof of concept (POC) using a quantum computer was developed with IBM’s Framework Opensource Qiskit, an infrastructure that included a simulator and IBM’s 16-qubit quantum computer. “In this sense, this technology did prove to be perfect for complex multivariate analysis,” he adds. This POC assessed the financial risk of two portfolios created specifically for the project based on real-world data, one consisting of mortgages and the other Treasury bills.
The bank has also worked with the PennyLane quantum computing framework from Xanadu, exploring the use of high-performance quantum random-number-generation solutions in simulations, alongside Quside Technologies, a Spanish spin-off of the Institute of Photonic Sciences.
In late 2021, a POC was started with VidaCaixa, the bank’s life-insurance and pensions arm, partnering with D-Wave Systems, which tested credit risk analysis using quantum computers. In this project, CaixaBank used the Leap Quantum Cloud Service to access D-Wave’s Quantum Hybrid Solver Service, which incorporates the Advantage quantum computer.
“This work usually requires an enormous amount of time and resources. However, when applying the quantum algorithm to the risk analysis, we reached the same conclusions as the traditional method, but in much less time cutting down the complex work of several days to just a few minutes,” says Mr Rebes.
Alternate uses
Research into the use of quantum computing goes beyond the world of trading and portfolio optimisation. Mastercard started its quantum journey in late 2019, says Steve Flinter, vice-president of AI and machine learning at Mastercard Foundry R&D.
“One of the things that we found out very quickly was, if you read any of the press and even academic papers that are looking at quantum computing in financial services, they tend to focus very much on problems in capital markets. So, things like derivatives, pricing, portfolio optimisation, and those kinds of areas — areas that are not terribly relevant to Mastercard and most of our customers,” he says.
Mr Flinter’s team started examining how quantum computing could be relevant to applications related to loyalty and rewards in the payments space.
“For example, if you’re running an offer or rewards programme — how can you decide which offers to give to which consumers or which card holders such so you’re maximising the benefit for everybody? [You want to be] maximising the value that your customers are getting, because they’re getting the best offers for them, and maximising the return on the investment that the business of running the rewards programme will be getting,” he says.
Mastercard, which is a networked business that connects banks, acquirers, card issuers, and merchants, also wanted to examine the challenges and problems inherent in how the payments company moved transactions around that network, he adds.
While no one interviewed for this feature described quantum computing as mature, Mr Flinter says that Mastercard guesses that a commercially viable quantum-powered application may be realised in financial services within “three to five years”.
“What I suspect we will find is that it’s not going to be a ‘big bang’, where suddenly you wake up on Sunday morning and computers can solve all the problems we want them to solve,” he says. Although, he admits that more difficult problems might take five, seven years or even more than a decade out to solve.
Experimental use cases
HSBC recently signed a three-year deal with IBM, allowing the bank to embark on a series of experiments to develop use cases for quantum computing in financial services.
As part of the relationship, HSBC will join the IBM Quantum Accelerator programme, giving it access to the tech firm’s quantum computing hardware, including its recently announced 127-qubit processor, Eagle, as well as quantum expertise, to help validate and progress potential quantum use cases.
According to Steve Suarez, global head of innovation, global functions at HSBC, the bank will explore the use of quantum computing for pricing and portfolio optimisation to advance its net-zero goals and mitigate risks, including identifying and addressing fraudulent activity. He goes on to say that the bank will upskill colleagues in quantum technology through internal training programmes, as well as actively recruiting quantum computing research scientists, to build a dedicated capability within its innovation team.
This agreement with IBM will be HSBC’s first formal foray into the burgeoning world of quantum computing. HSBC is also part of the European Next Applications of Quantum Computing project — a consortium of 12 European companies and research laboratories examining quantum use cases.
“The bet we’re making is that quantum advantage will be achieved relatively soon, and we will be able to upskill our people and be ready for it,” says Mr Suarez. “We want to make sure that once it’s ready, we’re ready with the right resources in place.” He admits that when quantum computing first started making inroads at various industries he was “a bit of a sceptic”, but now he believes that “what we want to do is be ready for it now, both from an opportunity perspective, and also from a risk perspective”.
In addition to the upskilling programme, HSBC plans on bringing in new people to “that are looking to really devote their careers to this”, he adds.
“I’m a firm believer that if we upskill our people, there’ll be a little bit more loyalty to stay with the bank,” says Mr Suarez. However, he does recognise that those new skills will make HSBC employees attractive to the industry. “We have to figure out how to retain that talent as much as we can and how to continue to bring in new talent,” he adds.
Interestingly, IBM seems to be the technology provider of choice for many financial services firms experimenting with quantum computing.
According to Katie Pizzolato, director of IBM Quantum Strategy and Applications Research, the last time many industries had a similar massive technology transformation was the 1950s, with the dawn of classical computing.
“We’re in a similar spot today with quantum,” she says. “But then we didn’t have the cloud computing, open-source software developer kits, and armies of computer scientists that we have available to us today to really push this technology much faster.”
Ms Pizzolato adds that IBM is running its quantum hardware roadmap in parallel with their software roadmap. “The workforce is going to have to develop in parallel as well,” she says. “I think that’s an important piece of this — that the barrier to entry is lower than most people think.”
