There are twelve half-a-million dollar refrigerators on the fourth floor of Yale’s Becton Center for Engineering and Applied Science. Cooled by liquid helium, they maintain a temperature of less than a hundredth of a degree Celsius above absolute zero. The refrigerators, which resemble industrial pipes, hiss in unison like a den of snakes. Inside each is a tangle of cords, wires, tubes, and copper plates inscribed with handwritten alphanumeric symbols: a primitive quantum computer.
Scientists and investors envision quantum computing as a technology exponentially more powerful than today’s classical computing. The speed and power of quantum computation has the potential to revolutionize the world of fast finance, create an unbreakable system of encryption, and model otherwise unpredictable atomic behaviors crucial to the development of chemicals and drugs. Recently, several major Silicon Valley firms have joined a growing number of academic groups in the race to create a working, marketable quantum computer. Quantum Circuits, Inc., a New Haven-based startup partnered with the Yale Quantum Institute, is considered a top contender to reach that goal, out of a field that includes IBM, Microsoft, and Google.
In January, I went to visit Luigi Frunzio, a senior member of Yale’s quantum computing department, who had offered to show me around. Frunzio has a bushy grey beard and a broad smile. He joined Yale’s team in 2003, during the lab’s infancy under the direction of professors Michel Devoret and Robert Schoelkopf. “It started at the beginning with each of them having one post-doc and a couple of PhD students,” Frunzio told me in his office. “Now each of them has five post-docs and twelve or thirteen graduate students.”
What makes quantum computers so much more powerful than classical computers is, most basically, a difference in their fundamental hardware. Classical computers store information on ‘bits’ that have a choice between two states: one or zero, on or off. In a quantum computer, on the other hand, two elementary units-—called ‘qubits’—are able to “communicate” with each other through the quantum phenomena of superposition and entanglement. Each qubit can be in two states at once, and together they can be in four. Adding more qubits increases the computer’s processing potential exponentially.
One of the tasks originally imagined for a quantum computer would be to break enormous numbers into prime factors in a matter of seconds. These problems can take today’s computers days or months to solve, and form the basis of modern file encryption and secure communications—Yale’s quantum computing lab is, in fact, almost entirely funded by federal defense agencies, including IARPA, the Intelligence Advanced Research Projects Activity. While it threatens to render current encryption methods obsolete, however, quantum technology also comes with its own, far more secure, form of encryption. Based on Heisenberg’s uncertainty principle—an axiom of quantum mechanics that states that it is impossible to measure a particle’s position without changing its momentum—quantum cryptography would protect communications from being intercepted. Since eavesdroppers cannot measure the characteristics of a quantum system without altering its state, they would be unable to avoid detection.
Frunzio explained that there remain many serious challenges to building a functional quantum computer, and even more to producing it on a commercial scale. For example, the qubits’ communications are only stable at incredibly low temperatures; hence the $500,000 refrigerators. “That already makes it a tool that my mom will never want to buy,” he said. He hopes, though, that the quantum world will eventually be downloaded to the “cloud” and available to everyone. IBM already put a simple five-qubit computer “online” a little over a year ago, and is planning to offer a sixteen-qubit version soon (though for a fee). Frunzio points out that, with so few qubits, the current online quantum computers are only really useful for abstruse calculations foreign to most people’s everyday life. But he imagines a not-so-distant future, in which users could be connected, for example, to “a video game on a quantum computer going so fast that it becomes like real life.”
In 2015, Frunzio, Schoelkopf, and Devoret spun off Quantum Circuits, backed by venture capital firms, with the goal of commercializing their research. The startup occupies an office on the second floor of 25 Science Park, a five-story concrete and glass cube at the edge of Dixwell, beyond Pauli Murray College. Yale owns the building and rents spaces to private companies; down the hall from QCI are Kleo Pharmaceuticals’ headquarters and a PepsiCo chemical lab. For now, QCI is entirely housed in one long room, the windowed side partitioned into office cubicles.
QCI is closely linked to the Yale quantum computing lab. All four of its employees are either graduates of Yale’s quantum physics doctorate program or worked there as postdocs, and, through a legal agreement with the university, the startup has exclusive rights to the Yale lab’s patents. However, all the researchers I met, both at Yale and at QCI, emphasized that the two projects have significantly different goals. Yvonne Gao, a graduate student studying with Professor Schoelkopf, explained that while the Yale academic efforts are focused on “the proof of principle experiments, demonstrating that certain principles of quantum mechanics could be realized in a laboratory experiment, QCI will have a very different emphasis, because for them it’s really about commercialization, the reproducibility of a certain type of device.”
QCI has already been successfully marketing one gadget for over a year: a quantum amplifier. “The amplifier is the read-out of the computer; it’s the way that you can read the qubits,” is how Frunzio described its function. Katrina Sliwa, who completed her Ph.D. under the direction of Professor Devoret in May 2016 and was hired that same month as QCI’s first employee, is an expert in quantum amplifiers. She explained that “to go from a proof-of-concept device to a product, you have to worry about how reproducible you can make these things, how reliable you can make these things, and then there are small changes to the design that can be made to make them more user-friendly.”
Professor Devoret helped to found QCI, but he recently left the company, feeling that tailoring quantum computing research to the market limits the scope of possible findings. Devoret, who is French, wears thick-framed glasses and speaks slowly and softly. “It’s not by perfecting candles that you discover electricity,” he told me in his office. “The industries only want to fund incremental research. They don’t want to fund wide research that looks far ahead in the future, they just want to make better products.” He prefers the way government grants are issued to Yale’s lab, less earmarked for specific projects and more designated for general exploration of the subject. “The great scientific discoveries have always been funded by governments. [The Yale lab’s money] is coming from the American taxpayer. And I think it’s healthy, I think that’s what the government is for, doing this long-term investment in knowledge.”
I asked Frunzio whether the first quantum computer to surpass classical capabilities will come out of Yale. “I don’t think it’s going to come out of the Yale academic effort. Because, you see, why would I ask one of my students to repeat the same experiment fifty or sixty times?” he said, adding, “But I hope that QCI will be in that race. Will it be the first? We’ll see.”
Frunzio is optimistic about QCI’s chances because he feels the competitive advantage in such a cutting-edge field lies primarily with small, focused startups, rather than enormous tech giants. Large companies are interested in staying competitive in relevant consumer markets and will only heavily invest in new domains once they become profitable. “It’s the difference between a mammal and a dinosaur,” Frunzio said. “You know, the dinosaur is big, but just to think to move his tail takes a long time because it’s so big.”
The mammal is certainly thinking hard. Its tail hasn’t moved much yet, but if you look closely, you’ll see it beginning to twitch.