- Goldman Sachs, RBS, CME, Guggenheim exploring the technology
- $10 million computers may help in complex asset allocation
You may soon be able to see every conceivable outcome of a wager, say on interest rates, over many time horizons -- thanks to computers that rely on the uncertainty at the root of all matter.
Wall Street is turning its attention to quantum computers, an emerging technology that aims to exploit the properties of subatomic particles to make extremely complex calculations at unprecedented speeds. Goldman Sachs Group Inc., Royal Bank of Scotland Group Plc, CME Group Inc. and Guggenheim Partners are evaluating quantum computing, which requires temperatures colder than in outer space.
The technology is exceptionally complicated, and researchers are debating whether machines now on the market are as powerful as advertised. If successful, these supercomputers could give money managers and banks, whose algorithmic-driven technology already dominate trading, another advantage in a highly competitive market. They can help managers better allocate money to a broad range of assets, find new ways to profit from differences in prices across markets, and value complex derivatives structures, said Marcos Lopez de Prado, a senior managing director at Guggenheim Partners.
“The quantum computer is very good at a few things that happen to be very, very hard for traditional computers,” said Lopez de Prado, also a research fellow at Lawrence Berkeley National Laboratory. “It can solve complex problems in exponentially less time.”
The cost: at least $10 million per computer, he said.
Financial firms are following a handful of companies and governments, including China, France and the U.K., into the world of quantum computing. Lockheed Martin Corp., one of the first adopters, has deployed it to spot bugs in the code that runs F-16 aircraft. Microsoft Corp. and International Business Machines Corp. are developing quantum technologies to speed up software. China’s Alibaba Group Holding Ltd. in July announced plans to form a quantum computing laboratory with the Chinese Academy of Sciences to carry out exploratory work.
Google Inc. is designing learning algorithms that work on a quantum computer and is collaborating in its research with the National Aeronautics and Space Administration, which has one of the machines. It has demonstrated the ability to crunch through very complex problems in a matter of seconds that would take traditional computers 10,000 years, Hartmut Neven, a director of engineering at Google, said at a press conference on Tuesday.
"We have encountered practical problems we would like to solve that are unfeasible with conventional computers," said John Giannandrea, Google’s vice president of engineering. "We want to understand the future that may lie ahead of us in non-conventional computers."
The quantum computers that can be bought today are made by D-Wave Systems Inc., a 16-year-old Canadian company. D-Wave has raised C$174 million ($130 million) from backers, including Goldman Sachs, which has invested its own money in the company.
At Goldman Sachs, engineers are considering using the technology, as is CME Group, the largest futures exchange, according to people who asked not to be identified because the information is private. Michael DuVally, a spokesman for Goldman, declined to comment. Alex Gorbokon, a spokeswoman for CME, said the exchange is “always looking at new technologies.”
Researchers including Guggenheim’s Lopez de Prado and Peter Carr, a Morgan Stanley managing director, said in a paper in August that they demonstrated the potential of quantum computing to help asset managers solve a difficult problem: wagering on a set of assets over a time horizon that’s divided into multiple steps. The manager must decide how much to invest in each asset at each step, while taking into account transaction and market-impact costs.
Managers today make their calls after traditional computers are fed historical data and forecasts for everything from interest rates and unemployment to stock and currency prices. Conventional technology is able to evaluate millions of possible investment scenarios to guide managers.
“That’s a drop in the ocean,” said Lopez de Prado. “Quantum technology can evaluate all possible scenarios in order to produce an optimized portfolio that can deliver the best risk-adjusted return.”
The performance of today’s binary computers is determined by the number of transistors that fit into a space and how cleverly engineers can connect them. Each transistor can essentially be switched on or off, which maps to zeros and ones -- or the bits -- in code that fuels the computer’s processor.
Quantum computers use qubits, which are based on measuring the properties of subatomic particles, as their most basic element. At any point in time, a qubit can be either a zero, a one, or somewhere in between. This means qubits in a quantum computer can explore a far larger number of possibilities than the zero-or-one bits in a typical computer to find the most efficient answer to a math problem, resulting in its greater power and speed.
For D-Wave’s computer to work properly and stop its qubits from interacting with other particles, its thumbnail-sized computer chip is encased in a large box, where it is cooled to temperatures of about minus 459 Fahrenheit (minus 273.15 Celsius).
CME is an investor in Canadian software firm 1QB Information Technologies, which helps companies write applications for quantum computers. 1QBit is working with a handful of financial institutions, including the Royal Bank of Scotland, said Andrew Fursman, co-founder of 1QBit.
"We’ve reached out to a number of large financial institutions and have a number of projects on the go," he said.
Simon McNamara, the chief administrative officer at RBS, told The Telegraph that quantum computing will play some role in finance, performing tasks in seconds that currently take days to complete.
D-Wave’s machines are spurring debate among researchers who question if they are truly quantum in operation. A paper published in 2014 found no evidence that D-Wave’s technology solved a task faster than a traditional computer using quantum methods. Then a paper published this summer found that the D-Wave machine displayed some quantum characteristics. That prompted Scott Aaronson, an associate professor at the Massachusetts Institute of Technology, to write that while researchers don’t fully understand the characteristics of its machine, "D-Wave and the others are in this for the long haul."
Lopez de Prado said the criticism is immaterial.
"I’m not so worried about how or why the computer solves the problem,” he said. “From a practical perspective what is clear is that this computer is able to do things that traditional computers can not.”
(An earlier version of this story corrected the name of 1QB Information Technologies.)