Behind Wells Fargo’s foray into quantum computing
Wells Fargo’s relatively new technology chief, Saul Van Beurden, is leading the company into quantum computing.
He recently led the bank to sign an agreement with IBM and MIT to collaborate on quantum computing and artificial intelligence technology. Van Beurden says the goal is to test and learn about the technology in order to determine its best use cases and potential impact on financial services.
“There are three camps in the industry,” said Van Beurden, a senior executive vice president and head of technology. “There’s the camp that says quantum will never come to production. There is the camp of people who believe it will take a long time before it will get in production, a long time defined as 10 or 15 years, maybe longer. And there is a camp that says, well, things might go faster. You’d better test and learn and be on it, because if this wave starts to take off, it might have game-changing impact for the industry. We are in the third camp. That's why we signed up. We don't want to be the bank that has regrets in a few years.”
Van Beurden left JPMorgan Chase as chief information officer of consumer and community banking to take the newly created role at Wells in April. He reports to new CEO Charlie Scharf.
Under the agreement, Wells Fargo will join the IBM Q Network, a community of large companies, startups, academic institutions and research labs working to advance quantum computing. IBM will provide Wells Fargo access to its quantum computing systems for use case exploration and fundamental research at the IBM Quantum Computation Center. And Wells Fargo will participate in the MIT-IBM Watson AI Lab, a collaborative project that IBM and MIT announced in September 2017, along with a $250 million investment from IBM.
It's just one of the most recent prominent moves by banks to dip their toes into quantum computing, a high-speed technology that takes advantage of quantum physics to solve complex problems.
NatWest has been experimenting with Fujitsu’s “quantum-inspired” Digital Annealer machine, which simulates quantum computing in silicon (Atos has a similar quantum computing simulator). The bank is running 1QBit software on the Fujitsu hardware to help portfolio managers decide on the right composition for the bank’s high-quality liquid assets portfolio. NatWest’s technology team has completed a portfolio allocation at 300 times the speed of a traditional computer, with a higher degree of accuracy, the bank says.
JPMorgan Chase and Barclays are also experimenting with quantum computing.
By 2022, 50% of global Tier 1 banks will use quantum computing to handle portfolio allocations, algorithmic trading and pricing strategies, according to IDC Financial Insights. This prediction includes the use of quantum computer simulators like those from Fujitsu and Atos, as well as actual quantum computers from companies like Google, IBM and D-Wave.
Jerry Silva, research director for IDC Financial Insights, said this technology will be largely the domain of big Wall Street banks because it's so expensive.
“I see the top-tier institutions only using it to run highly critical applications,” he said. “If you look at trading, for example, that's a perfect opportunity to gain competitive advantage if you can make a decision that much faster than your competitor. This is going to be a huge differentiating factor for things like trading and portfolio balancing.”
Van Beurden said he has no defined timetables for this project.
“We see this as research and development work where we invest without a predictable timeline or outcome,” he said.
He expects to use quantum computing to analyze risk or calculate prices.
“But I am sure that once you start doing this, you will find other use cases that we are not thinking of yet, just because of the fact that you will bring different people with different backgrounds together in the research area,” Van Beurden said. “The nice thing is, I don't know yet what the use cases will be.”
Wells Fargo will contribute data scientists and Ph.D.s in quantum computing to the MIT-IBM lab.
“We will chime in with a small team of the best and the brightest in those areas where we have the first hypothesis and the use cases in mind,” Van Beurden said. “That's often the biggest chance for success rather than the large team that starts to spin in all directions. It's great to have a small team working together with a small team of IBM and other research institutes to find out what are the use cases for the banking industry that we can leverage together.”
Bob Sutor, vice president of IBM ecosystem development, estimates that by the end of the next decade, quantum computers will be large enough and powerful enough to begin solving practical problems that customers like banks care about significantly better than traditional computers.
“One of the things that Wells Fargo’s participation represents, along with the 80-plus other members of the Q network, is that many people think this is serious and are willing to invest the time and the money to speed up this process,” he said. “In three to five years, we think it's likely we'll see some classical problems get done using quantum computing.”
The sweet spots for quantum computing in financial services will be risk analysis, derivatives pricing and settlements, he said.
“The essence of those problems is that, while they're not very hard to describe, and in fact they don't even use that much data relatively speaking, they involve many possible combinations of different possibilities,” Sutor said.