Vern Brownell, who for eleven years was chief technology officer at Goldman Sachs, is excited about the potential for quantum computing in banking.
"I saw the best of all the new technology there was, but I never saw anything that could potentially change computing dramatically in the entire time I was at Goldman" the way quantum computing could, says Brownell, who is now CEO of Vancouver-based quantum computing company D-Wave.
Quantum computing directly leverages quantum mechanics, the laws of physics that govern the smallest particles in the universe, to solve problems at high speeds. Traditional computers only allow bits of information to live in one state (0 or 1) at a time. A quantum computer, operating in an extreme environment, uses qubits (quantum bits) that enable bits of information to be a 1, 0 or both 0 and 1 simultaneously. The result is a computation system that can manipulate and assess many combinations of information concurrently.
A quantum computer can cycle through 10 to the 154th power potential answers to a problem in microseconds. "By way of reference, there are only 10 to the 80th power atoms in the universe," Brownell says.
In tests, he says D-Wave's machine has performed 100,000 times faster than high-end Intel cores.
D-Wave makes quantum computing processors, a 10-meter-square "rarified environment" that houses the chips, and compilers and software that make it all usable by regular software applications.
The processor, which is the size of a human fingernail, is not unlike an Intel chip in some respects. "But it's a superconducting chip, so it runs at near absolute zero [temperature]," he says. "There's a lowest possible temperature in physics that you can't go below, and we're .01 degrees above that, so we're almost at the very lowest temperature you can have in the universe. In fact, we're 150 times colder than interstellar space."
D-Wave's enclosure operates somewhat like a common refrigerator, except it uses liquid helium as a coolant rather than Freon.
In addition to extreme cold, the quantum bits require a magnetic vacuum; the enclosure removes the magnetic field as well.
This all sounds rather high-maintenance, but Brownell says the machine doesn't consume an excessive amount of energy.
"I'm an old data center guy and I spent a lot of time building out these large, power-hungry data centers," he says. "One of the interesting things is at these very low temperatures, when you put a current through a wire, there is no resistance so it generates absolutely no heat. Today we have 512 qubits, we could have a billion qubits, and it will be the same amount of heat — zero."
All told, the equipment consumes about 15.5 kilowatts, which is amount the same as a few servers, Brownell says.
Using the interfaces D-Wave has built, the machine can be used to run Java, Python or C++ programs, Brownell says.
"It's a cloud service and you can access it across the network and integrate it with traditional computing platforms," he says.
The sweet spot for this technology is mathematically intense problems that scale exponentially as complexity grows, Brownell says. In financial services, Monte Carlo simulations for risk and/or pricing purposes are one example — today, many banks run such simulations on clusters of servers running thousands of core processors. The quantum computer could function as a high-performance coprocessor in a traditional computing environment, the way Wall Street firms for a time turned to field programmable gate arrays and graphics processor units to run their pricing models.
Quantum computers could also be used for portfolio optimization, to find the best mix of assets for a client. They could potentially be used anywhere data analysis or machine learning is needed, including customer, marketing, fraud or risk analytics.
D-Wave is working with a few partners in the financial services industry to develop packaged solutions for the industry, although Brownell won't specify which ones.
"We're working with companies at the top of their game in financial services," he says.
D-Wave has received investments from Goldman Sachs, Amazon founder Jeff Bezos, and In-Q-Tel, an investment firm with close ties to the CIA and other government agencies.
The company has gone public on its work with Lockheed Martin, NASA and Google. Lockheed Martin uses the quantum machine to test and validate flight control systems. NASA is using it to help search for new planets that might harbor life outside our solar system. Google is exploring using it to build more accurate models for speech recognition, web search and bioinformatics.
It's still early days for quantum computing. "If you look at the dawn of microprocessors, there weren't compilers and tools for those — Bill Gates and Paul Allen built those initial compilers," Brownell says. "We're in the process of building a lot of those software tools that make it easier to use."