Imagine a Wall Street where complex options probability models are calculated instantly and simulations based on massive data sets run with ease. At JPMorgan, that future may not be too far off, according to execs who have overseen the bank's foray into quantum computing.
As an emerging alternative to the traditional computer, JPMorgan execs believe quantum computing could upend how finance firms perform computations. It's why the largest US bank by assets is experimenting with quantum computing through a group it calls Global Technology Applied Research.
Unlike classical computers, which can store only a one or a zero, quantum computers use quantum bits, or qubits. Qubits can store multiple values at the same time, which theoretically gives the quantum computers big speed advantages.
It's a bet that stands to pay off handsomely by helping the bank execute calculations milliseconds faster than the competition in functions from risk analytics to algorithmic trading.
The ability to use quantum computing for practical applications is still largely theoretical, and critics say the hardware is not yet advanced enough and that quantum machines require more resources to run than is worthwhile. But JPMorgan researchers say there are signs the industry is getting close to the point where quantum computers can process real-world problems faster than classic computers, often referred to as quantum advantage.
"We are still some years away from quantum advantage or quantum supremacy, but there has been progress in hardware, and we are past theoretical conversations now," Marco Pistoia, the head of global technology applied research, told Insider. "It's a good time to experiment so that we can be quantum-ready."
Quantum computing meets artificial intelligence
Global Technology Applied Research was created in 2020 and is led by Pistoia and Andrew Lang, the bank's global chief technology officer. Some use cases the bank is exploring are portfolio optimization, option pricing, risk analysis, and fraud detection and natural language processing, Pistoia said.
The group works closely with the bank's artificial-intelligence and machine-learning teams, Lang told Insider via e-mail.
"There are also numerous similarities from a software-platform perspective because the best way to reap the benefits of quantum computing or AI/ML is a cloud-based hardware-agnostic framework that allows different teams of researchers and developers to share the same programming environment," Lang said.
Machine-learning and AI problems tend to exhibit high complexity, making them prime candidates for quantum's ability to process computations at breakneck speed. For example, a 2019 research blog from IBM, which has conducted quantum research with JPMorgan, said the bank could execute an algorithm used to predict options prices in nearly real time, as opposed to overnight. Plus, many of the bank's quantum algorithms have AI and machine-learning logic embedded in them.
The shared platform, overseen by Castillo, is also hardware-agnostic, meaning it can interface with quantum computers from different providers without needing to rewrite the application code.
Wall Street's race for quantum is heating up
Investing in quantum computing isn't without risks. This August, the Financial Times published an opinion piece written by Oxford University physicist Nikita Gourianov. The core of Gourianov's argument is that both the future promise and current applications of quantum computing are overstated — and that too many dollars are chasing little productive gains.
But JPMorgan is far from the only Wall Street firm betting that these investments will prove rewarding. Goldman Sachs' research and development engineering team is developing tools to more quickly assess markets and refine how it prices some financial instruments. The bank has also invested in budding quantum startups, such as QC Ware, which develops code bases that include quantum-processing elements that can run on any of the four main public-cloud providers.
Meanwhile, Fidelity has been exploring the potential of quantum computing in wealth management. Together with Amazon Web Services, Fidelity has built an algorithm that mimics the behavior of a market index, which Fidelity then used as a training ground to teach employees how to think through quantum algos and write programs for quantum computers.
Wells Fargo, too, has invested in quantum computing to process complex data structures used for fraud detection. Quantum is expected to reduce the weeks-long time period it takes for new customers to be onboarded.
