As Geordie Rose points out, you can’t have a conversation with your laptop. “It’s worth stopping and thinking about why that is,” says the founder and chief technology officer of D-Wave Systems Inc., which sells what it says is the first commercial quantum computer.
Part of the reason is that conversation is nonlinear, so such a feat would require fairly sophisticated artificial intelligence (AI).
“With these types of [quantum] computers, there’s a very clear path toward resolving that problem,” says Dr. Rose in a boardroom at D-Wave’s modest headquarters in the Vancouver suburb of Burnaby, B.C. He’s referring to the fact that quantum computing promises to get computers nearer to the kind of complex processing the human brain can do and that AI hasn’t yet been able to replicate. This nascent technology can handle information not just in binary format (zeros or ones) but harness the power of quantum mechanics to deploy zeros and ones at the same time. It has the potential to be millions of times more powerful than today’s supercomputers, solving complex problems in minutes that currently would take years.
“All we need is more hardware,” Dr. Rose says.
In the development lab just across the hall, D-Wave has plenty of hardware on display. One corner of the cluttered space contains four of its supercooled systems, including a new model packaged for sale in a three-metre-tall black box that serves as a shielded room to house the computer.
Although it has detractors, D-Wave is a world leader. No other company has come close to building a similar machine, let alone sold one. Quantum computing could spawn a new industry as big as the conventional silicon-based computer business, Dr. Rose and other proponents say. D-Wave owes much of its success to an innovative early decision to outsource research and development.
Soon after the company launched in 1999, it created an international research network that tapped some of the world’s top experts in quantum computing. This exercise helped D-Wave quickly determine what kind of quantum machine to build so it could make a product. It also locked down patents that puts would-be rivals at a big disadvantage.
Not everyone thinks D-Wave is the real deal. Since it unveiled its first working computer in 2007, the 72-employee Canadian company has faced skepticism from purists, who say the D-Wave system is a pale imitation whose circuitry doesn’t obey the laws of quantum physics. In its defence, D-Wave cites a 2011 paper in the reputable scientific journal Nature as proof that quantum properties are in play.
And as an entrepreneur, Dr. Rose has support where it counts. Investors, including prominent U.S. venture capital firms, have staked $130-million on the Montreal-raised theoretical physicist’s big idea. D-Wave is beginning to attract paying customers, too. In 2010 it sold its first system, to aerospace and defence giant Lockheed Martin. And last month it closed $30-million in equity financing from two big names: Bezos Expeditions, a venture capital shop established by Amazon.com Inc. founder Jeff Bezos; and In-Q-Tel, which sources technology for U.S. intelligence agencies.
“They, a decade ago and today, are the only company trying to make a quantum computer,” says venture capitalist Steve Jurvetson, managing director at Menlo Park, Calif.-based Draper Fisher Jurvetson, a D-Wave investor since 2003.
The company could be on the verge of unleashing vast computing power. Quantum computers handle information in a fundamentally different way than so-called classical computers. A D-Wave processor doubles in power every time its developers add a quantum bit, or qubit, a basic building block that is the equivalent of transistors in classical silicon chips. As it prepares to launch a 512-qubit product before the end of 2012, the company has proven that it can roughly double the number of qubits every year.
“A quantum computer is on a completely different scaling curve, where once it passes traditional computers, they can never catch up,” says Mr. Jurvetson, who sits on D-Wave’s board. “That is just unprecedented in the technology business.”
But D-Wave isn’t building a general-purpose quantum computer, he adds. “It does one very specific class of problems really, really well, and that’s pretty much all it does,” he says. “But luckily that class of problems has wide applicability.”
For example, it could eventually use a person’s genome to determine how they will respond to a particular drug, Dr. Rose says. “Virtually anything that a human does well that conventional computers currently are not good at is something that’s going to be affected by these systems.”
So far, quantum computers are only starting to deliver on that promise. In 2009, for instance, scientists at Google Inc. used a D-Wave system to accurately detect cars in images. Dr. Rose describes such automatic detection as a fundamental problem in AI. “In some ways, understanding how we do that is the key to unlocking intelligence in machines,” he says.
When D-Wave launched, the field of quantum computing was largely theoretical, Dr. Rose recalls. “It was very early, but it wasn’t so early that the science hadn’t been in some ways proven out,” he says. “There were enough results that had been garnered from the scientific study of these things to think that there was nothing written in the laws of physics that prevented you from trying to build one.” So he and his colleagues mapped out the couple of dozen areas that someone would need to understand in order to build a real quantum computer, from the user experience to the physical devices inside the chips. Then they looked for scientists around the world with expertise in those areas and asked if they would be interested in helping D-Wave.
The company gave its researchers funding and access to the rest of the network it was building. In exchange, it got control of the intellectual property they produced and the right to file patents before they published their findings.
Over the next five years, D-Wave’s network expanded to include groups with ties to 10 academic institutions in Canada, Germany, the Netherlands, the Slovak Republic, Sweden, Britain and the Ukraine. As early as 2001, the start-up had access to $440-million worth of equipment.
D-Wave took an entrepreneurial approach to running tests that would otherwise have been very expensive, says Ajay Agrawal, Peter Munk Professor of Entrepreneurship at the University of Toronto’s Rotman School of Management. “If you had all of the equipment in-house, you’d have to be a very large company, like an IBM,” explains Dr. Agrawal, who co-authored a 2004 Harvard Business School case study on D-Wave. “So how does an entrepreneur do it? By leveraging the assets that are out there and in many cases underutilized, and rather than paying the full cost, paying only the marginal cost.”
D-Wave handled what was probably the riskiest stage of its life in a surprisingly careful and efficient way, says Alexei Andreev, executive vice-president and managing director at the Palo Alto, Calif., office of venture capital firm Harris & Harris Group Inc., another investor. “Instead of trying things sequentially, they did it in parallel,” the D-Wave board member says.
By 2003, D-Wave had assembled a portfolio of knowledge and patents that was unrivaled in the quantum computing field, Dr. Rose says. At a very early stage, the company focused on protecting its intellectual property through patents, notes director Geoff Catherwood, a Vancouver-based partner with BDC Venture Capital’s Energy/Clean Tech fund. “When we invested, they had under 20 employees, but they had three people dedicated to IP,” says Mr. Catherwood, whose firm closed its first D-Wave investment in 2002.
After weighing all of its options, D-Wave started to build the computer. Drawing on its findings from the research network, it settled on a particular model of quantum computing for the product it went on to develop in Burnaby. “Now we have this thing that has been remarkably successful at solving the kinds of problems it’s designed to solve, and the markets that it addresses are huge,” Dr. Rose says. “So I think that the guess that we made at the end of the collection of all of this information was the right one.”
From an investment standpoint, the network also brought D-Wave scientific credibility, Dr. Agrawal says. “When investors wanted to do due diligence, they could go and openly talk to researchers in the network.”
Today, D-Wave holds 93 U.S. patents and has 107 patent applications under way globally. Its IP portfolio will make it very difficult for competitors to design a similar machine, at least for 15 years or so, Dr. Rose predicts.
D-Wave’s next challenge is how to build the company in a way that maximizes value, Harris & Harris’s Dr. Andreev says. “I’ve never seen an opportunity with this sort of unlimited upside. Now the question is how to play it properly.”
Quantum Computing 101
A traditional computer processes information as bits that can be a 0 or a 1. A quantum computer exploits the laws of quantum physics by making its bits a 0, a 1, or a 0 and a 1 simultaneously. This “superposition” lets it do many calculations at once, where a traditional computer can only perform one
The most popular approach to building a quantum computer is the circuit or gate model, whose processor architecture resembles that of conventional computers.
D-Wave Systems Inc. uses the relatively new adiabatic model, also known as quantum annealing. This architecture allows its quantum bits, or qubits, to shift from superposition to a traditional computer state.
D-Wave's processor circuitry is made from the metal niobium, which turns into a superconductor at very low temperatures, so the processor is supercooled to just above 0º Kelvin (-273.15° Celsius). The D-Wave processor resides in a cylindrical refrigerator suspended inside a shielded room, with 16 layers protecting it against everything from radio-frequency noise to magnetic disturbances.
The processor consists of qubits connected by couplers; surrounding them is a programmable magnetic memory.
Since demonstrating what it called “the world’s first commercially viable quantum computer” in 2007, D-Wave says it has progressed from a 16-qubit to a 512-qubit processor. While this shows an ability to increase the quantum computing power of the computer, it is not clear yet what real-world applications that kind of processor has.
Join the conversation on Canada's competitiveness by following Canada Competes on Twitter:@CanadaCompetes