Geordie Rose wasn’t sure it would work.
It was February 2013, and Rose was waiting for Catherine McGeogh, an Amherst University professor in computer science, to run the third of a series of trials to prove that his company’s latest computer model, the D-Wave Two, could solve some of the world’s most fiendish computational problems. His potential clients—Google, NASA, and the Universities Space Research Association (USRA), a global consortium of universities and government research centres—are some of the most demanding in the world, and they wanted to put the D-Wave Two through its paces before committing to buy. A good deal of D-Wave’s scientific credibility—and the company’s financial fate—hinged on the outcome of this test.
At the other end of the room sat the computer itself: an enigmatic glossy black cube, 10 feet across on each side. Beneath its slick exterior is a shielded chamber to protect the inner workings from noise, vibration and magnetic interference. A series of refrigeration units cool the chamber with liquid helium. Inside this vacuum-sealed container, a stacked series of five gold-plated copper plates hangs down, separated by wiring and more cooling components; each plate is progressively colder, leading down to a coffee-can-sized cylinder that houses the D-Wave’s thumbnail-sized processor, which must be kept at 0.01 degree kelvin, or –273°C. In the room outside, the refrigeration equipment emits a constant low humming, and the occasional high-pitched chirp courtesy of pulsing helium gas within the cryocooler. Despite all of this, the whole machine runs on just 10 kilowatts of power—even less than a standard refrigerator.
The D-Wave Two isn’t just faster than other computers; it’s the first commercially available quantum computer, which uses the fundamental principles of physics to solve problems in a completely different way. By exploiting the strange mathematical properties that reign at the subatomic level, D-Wave’s chip is theoretically able to do more calculations, faster, than any comparable computer chip that currently exists. If it works, it stands to revolutionize computing; but getting the concept off the drawing board is no easy matter.
Raymond Laflamme, director of the University of Waterloo’s Institute of Quantum Computing, says we are on the cusp of another computing revolution, with quantum computing enabling everything from more accurate weather forecasts to advanced artificial intelligence. “Today, we’re looking at a small piece of a door that is wide open,” he says. “Will the impact of quantum computing equal that of the advancements of the last 50 years? My gut feeling is yes.”
On this particular occasion, the D-Wave Two would have to pass five somewhat more mundane tests, pitting it against a variety of conventional processors. The exact nature of those tests is still secret, but the companies could confirm they were similar to common theoretical computing problems such as the “travelling salesperson” test. In that scenario, a computer is given a list of cities and the distances between them; the challenge is to map the shortest possible route that visits each city just once and then returns home. It’s a kind of mathematical optimization problem that can be applied to all sorts of other computing tasks: logistics, pattern recognition, cryptography and more.
For Rose, a $15-million sale, and perhaps the future of the company, depended on the outcome. Now all he could do was wait.
In late April, Mike Lazaridis stood on a small auditorium stage at the University of Waterloo’s new Mike & Ophelia Lazaridis Quantum-Nano Centre to talk about why, over the past 12 years, he has invested more than $300 million in an attempt to establish Waterloo, Ont., as the world capital of quantum computing. The audience was mostly composed of early-stage VCs and other potential investors, and Lazaridis’s primary message to them was that even though quantum computing is in its infancy, they need to jump on the opportunity now.
“The work that we’ve done strategically over the past 12 years has given this region a unique opportunity to be one of a handful of global centres that will lead the second quantum technology revolution,” Lazaridis said. “It is essential that we work together to ensure Canada continues to play a meaningful role. So what do we need to do?”
Over the past decade, researchers around the world have been trying to answer that question. Lazaridis is determined that Canada should take a leadership role in quantum technology. The Research In Motion founder has emerged as one of quantum’s most enthusiastic evangelists, and he’s wagered a considerable chunk of his personal fortune on its future. In March, just a few weeks before this appearance in Waterloo, he and his BlackBerry co-founder Doug Fregin launched the $100-million Quantum Valley Investment Fund to commercialize spinoff technologies in quantum science.
“How do you know what to invest in?” Lazaridis challenged from the stage. “The best example I can give is to invest in something that nobody understands. Because if you understand it, you’re probably too late.”
Quantum computing defies understanding in part because it’s not just insanely complicated; it also doesn’t obey any of the laws of physics we encounter in our daily lives.
Classical physics—the kind we know about courtesy of Galileo and Newton—is comparatively easy to understand because we can clearly see it working all around us: the apple falls from the tree; the earth orbits the sun; the thrown baseball follows an arc that we can predict with an equation.
But at the subatomic level, quantum mechanics takes over, and that’s when things get weird: electrons appear to be in two places at the same time; just looking at a photon changes its behaviour; particles can behave as if they’re telepathically linked, mirroring each other’s actions instantly even though they’re far apart. Much of this is still the subject of lively scientific debate. But the strangeness of the quantum world also offers some exciting possibilities.
In a traditional computer, all information is encoded as binary bits: every bit is a one or a zero. If you can shuffle enough of them around fast enough, you can solve complex equations, which has changed the way we do everything from trading stocks to simulating nuclear explosions to storing recipes.
A quantum computer instead uses quantum bits—called qubits—which are atomic-scale structures that, through a phenomenon known as “superposition,” can be both zero and one at the same time. That strange quality of quantum mechanics allows qubits to complete difficult calculations many times faster than traditional bits, which have to be one or zero, on or off, black or white.
Say you’re standing in a very large room, all the lights are off, and you have to find the only exit. If you’re a conventional computer, you would walk to the outside wall and feel along, inch by inch, until you found the door. But if you’re a quantum computer, a thousand versions of you search simultaneously; the one that finds the exit opens the door, and the other 999 simply vanish.
If you can find a way to harness that power on a chip, this approach can be used to run a massive number of complicated simulations and models much quicker than we’ve ever been able to before.
“This type of computer is not intended for surfing the Internet, but it does solve this narrow but important type of problem really, really fast,” says McGeogh, who designed the tests used on the D-Wave machine.
The ability of quantum computers to run travelling-salesperson–type problems at lightning speeds can help make more accurate predictions much faster—whether that’s figuring out how certain diseases develop, how our climate is changing or better understanding your speech to find what you’re looking for on the web.
But even the scientists in the field are limited in what they can predict. Laflamme compares it to how people in the 1950s described the then-dawning age of computers. Spoiler alert: no one predicted the iPhone.
Computer scientists are only beginning to figure out how they might apply this kind of computing power. To get an idea of the brain-bending scale these machines can operate at, consider this: a quantum computer with 300 qubits working could run more calculations in the blink of an eye than there are atoms in the entire universe—and the D-Wave Two processor has 500 qubits.
Despite the potential, quantum computing is still a hard sell for most investors, thanks to the potential 20- or 30-year timelines involved. OMERS Ventures’ David Crow was in the audience for Lazaridis’s Waterloo talk and says it’s too early for most VCs to get involved. “[Venture capital funds] tend to have a seven- to 10-year lifecycle, so a lot of it is outside the scope of a single fund,” he says. “There’s a huge amount of commercialization that needs to happen before you can start building companies on it.”
Rose says that traditional investment approach is changing. “The people who are at the vanguard of the investment community in and around San Francisco are starting to come to grips with the fact that there’s a gaping hole in the strategy that investors are using today, and are now looking at things that are disruptively world-changing, fundamental technologies that will take five to 15 years to develop and are extremely capital intensive,” he says.
Between Lazaridis’s Quantum Valley and Rose’s D-Wave, Canada has two very different but no less significant approaches to this coming revolution.
First, there is the Waterloo way: funding research through the university and using that as the fertile ground to cultivate a healthy startup ecosystem. Second is D-Wave’s approach, which is to use the company as the R&D base, tapping large-scale clients like Fortune 500 companies and governments that have the size and scale to make large up-front investments.
D-Wave has been criticized over the years by some academics who say its machines aren’t true quantum computers, but high-performance conventional computers with some quantum characteristics. That hasn’t bothered Rose, who is quick to admit the D-Wave Two is just a very, very early step on the road toward the true potential of quantum computing. But he also says he’d rather build something than split hairs with academics. D-Wave would be happy to sell you a quantum computer, today, if you have a few spare million lying around and don’t mind having a slightly sinister-looking, 10-foot-tall obsidian cube humming away in a corner, occasionally chirping. You’ll be in good company: when Google, NASA, and USRA finished kicking the tires, they banded together and bought one of their own.
Yes, the D-Wave Two passed its five secret tests—with flying colours. “If you want it to solve the exact problem it’s built to solve, at the problem sizes I tested, it’s thousands of times faster than anything I’m aware of,” says Catherine McGeogh, the professor that Google and company brought in to act as an independent adviser on their tests. In a business known for glorifying speed, quantum computing promises to—in certain circumstances, at least—change the scale on which we measure it. (McGeogh cautions that comparing the D-Wave to conventional computers with traditional metrics isn’t even apples to oranges—more like apples to fish.)
Rose is less moderate. “No computing technology, ever in the history of mankind, has required the kind of speed they asked us for,” says Rose. “And we blew them out of the water.”
And so in May, Google, USRA and NASA announced D-Wave’s $15-million machine would take up residence at their new Quantum Artificial Intelligence Lab in Silicon Valley to look at potential breakthroughs in artificial intelligence.
That deal was announced barely a month after D-Wave announced a $10-million deal with Lockheed Martin. Last fall, a collection of investors, including Amazon CEO Jeff Bezos and the Central Intelligence Agency’s investment arm invested $30 million in D-Wave, and on Sept. 20, the company cut the ribbon on a new 15,000-square-foot laboratory and office facility in Burnaby, B.C., that doubles its capacity.
Watch this interview with D-Wave CEO Vern Brownell from the Canadian Business Leadership Forum:
The deals that D-Wave is ramping up are significant signs that this technology is finally making its way from the lab bench to the loading dock. It’s also a sign that the race to commercialization has officially begun.
Lazaridis seems giddy about the commercial potential. “There’s a passage in the Bible that talks about…when someone finds a pearl of great price in a field, he sells all of his possessions and buys the field,” he told the Waterloo audience. “This is a truly profound moment in our history.”
Developments like the D-Wave Two and Quantum Valley Investments have changed the question from whether Canada will have a hand in the next computing revolution, to just how big its contribution can be. The momentum in Waterloo and Burnaby is something to be proud of, but both Rose and Lazaridis are quick to caution that these are early days, and IQC’s Laflamme warns that Canada has to keep up its momentum in the field. “We are leaders now, but we can’t slow down or we risk falling behind,” he says. “There are plenty of others around the world pushing these boundaries—Europe, the U.S., China—so we need to keep going.”
Canada has long had success building prototypes and early-stage businesses, only to lose them to foreign acquisition or investment. As the country struggles to find its innovation mojo outside the resource-based economy, there needs to be a strategy among all levels of government and investors to help leaders like Lazaridis and Rose, and the entrepreneurs who grow as a result of their efforts.
Rose hopes D-Wave’s recent sales can be the spark that ignites the quantum revolution for Vancouver, saying there will be opportunities for others in the wake of his company’s success. “We’re primarily a hardware company, but there are opportunities for companies building software where they could become the next Microsoft,” says Rose. “It could happen here, or it could happen in California. I’d prefer it happen here.”