Hoping for quantum leap
Yale professors’ startup in computing races Google, IBM and Intel
Robert Schoelkopf is at the forefront of a worldwide effort to build the world’s first quantum computer. Such a machine, if it can be built, would use the seemingly magical principles of quantum mechanics to solve problems today’s computers never could.
Google, IBM, and Intel are using a method pioneered by Schoelkopf, a Yale University professor, and a handful of other physicists as they race to build a machine that could significantly accelerate everything from drug discovery to artificial intelligence. So does a Silicon Valley startup called Rigetti Computing. And though it has remained under the radar until now, those four quantum projects have another notable competitor: Robert Schoelkopf.
After their research helped fuel the work of so many others, Schoelkopf and two other Yale professors have started their own quantum computing company, Quantum Circuits.
Based just down the road from Yale in New Haven, Conn., and backed by $18 million from Sequoia Capital and others, the startup is another sign that quantum computing — for decades a distant dream of the world’s computer scientists — is edging closer to reality.
“In the last few years, it has become apparent to us and others around the world that we know enough about this that we can build a working system,” Schoelkopf said. “This is a technology that
we can begin to commercialize.”
Quantum computing systems are difficult to understand because they do not behave like the everyday world we live in. But this counterintuitive behavior is what allows them to perform calculations at rates that would not be possible on a typical computer.
Today’s computers store information as “bits,” with each transistor holding either a 1 or a 0. But thanks to something called the superposition principle — behavior exhibited by subatomic particles like electrons and photons, the fundamental particles of light — a quantum bit, or “qubit,” can store a 1 and a 0 at the same time. This means two qubits can hold four values at once. As you expand the number of qubits, the machine becomes exponentially more powerful.
Todd Holmdahl, who oversees the quantum project at Microsoft, said he envisioned a quantum computer as something that could instantly find its way through a maze. “A typical computer will try one path and get blocked and then try another and another and another,” he said. “A quantum computer can try all paths at the same time.”
The trouble is that storing information in a quantum system for more than a short amount of time is very difficult, and this short “coherence time” leads to errors in calculations. But over the past two decades, Schoelkopf and other physicists have worked to solve this problem using what are called superconducting circuits. They have built qubits from materials that exhibit quantum properties when cooled to extremely low temperatures.
With this technique, they have shown that, every three years or so, they can improve coherence times by a factor of 10. This is known as Schoelkopf ’s Law, a playful ode to Moore’s Law, the rule that says the number of transistors on computer chips will double every two years.
“Schoelkopf ’s Law started as a joke, but now we use it in many of our research papers,” said Isaac Chuang, a professor at the Massachusetts Institute of Technology. “No one expected this would be possible, but the improvement has been exponential.”
These superconducting circuits have become the primary area of quantum computing research across the industry. One of Schoelkopf ’s former students now leads the quantum computing program at IBM. The founder of Rigetti Computing studied with Michel Devoret, one of the other Yale professors behind Quantum Circuits.
In recent months, after grabbing a team of top researchers from UC Santa Barbara, Google indicated it is on the verge of using this method to build a machine that can achieve “quantum supremacy” — when a quantum machine performs a task that would be impossible on your laptop or any other machine that obeys the laws of classical physics.
There are other areas of research that show promise. Microsoft, for example, is betting on particles known as anyons. But superconducting circuits appear likely to be the first systems that will bear real fruit.
The belief is that quantum machines will eventually analyze the interactions between physical molecules with a precision that is not possible today, something that could radically accelerate the development of new medications.
Google and others also believe that these systems can significantly accelerate machine learning, the field of teaching computers to learn tasks on their own by analyzing data or experiments with certain behavior.
A quantum computer could also be able to break the encryption algorithms that guard the world’s most sensitive corporate and government data. With so much at stake, it is no surprise that so many companies are betting on this technology, including startups like Quantum Circuits.
The deck is stacked against the smaller players, because the big-name companies have so much more money to throw at the problem. But startups have their own advantages, even in such a complex and expensive area of research.
“Small teams of exceptional people can do exceptional things,” said Bill Coughran, who helped oversee the creation of Google’s vast Internet infrastructure and is now investing in Schoelkopf ’s company as a partner at Sequoia. “I have yet to see large teams inside big companies doing anything tremendously innovative.”
Though Quantum Circuits is using the same quantum method as its bigger competitors, Schoelkopf argued that his company has an edge because it is tackling the problem differently. Rather than building one large quantum machine, it is constructing a series of tiny machines that can be networked. He said this will make it easier to correct errors in quantum calculations — one of the main difficulties in building one of these complex machines.
But each of the big companies insist that they hold an advantage — and each is loudly trumpeting its progress, even if a working machine is still years away.
Coughran said that he and Sequoia envision Quantum Circuits evolving into a company that can deliver quantum computing to any business or researcher that needs it. Another investor, Canaan’s Brendan Dickinson, said that if a company like this develops a viable quantum machine, it will become a prime acquisition target.
“The promise of a large quantum computer is incredibly powerful,” Dickinson said. “It will solve problems we can’t even imagine right now.”
“This is a technology that we can begin to commercialize.” Yale Professor Robert Schoelkopf