All products featured on Wired are independently selected by our editors. However, we may receive compensation from retailers and/or from purchases of products through these links.
Within the next five years, Google will produce a viable quantum computer. That's the stake the company has just planted. In the pages of Nature late last week, researchers from Google's Quantum AI Laboratory told the world that a machine leveraging the seemingly magical principles of quantum mechanics will soon outperform traditional computers on certain tasks. They said this long-anticipated technology will, among other things, improve the artificial intelligence that's already remaking the tech world. "The field of quantum computing will soon achieve a historic milestone," the team wrote. They call this milestone "quantum supremacy."
Now IBM is planting a stake of its own. Today, the company announced plans to offer commercial quantum machines to businesses and research organizations within the year. These machines will not bring quantum supremacy—namely, they won't surpass the performance of traditional machines. But much like Google, IBM claims it will reach that threshold over the next few years. "We are reaching a key moment," says IBM research vice president Dario Gill.
It's no accident that these two announcements arrived at about the same time. A true quantum computer is not yet a reality. "You can't do anything practical today," says Gregoire Ribordy, founder and CEO of quantum cyber-security company ID Quantique. But the world's biggest tech companies are already jockeying for their own form of commercial supremacy as they anticipate a quantum breakthrough. Both Google and IBM now say they will offer access to true quantum computing over the internet (call it quantum cloud computing). Microsoft recently hired several notable researchers in launching its own effort to build a quantum computer. And in China, internet giant Alibaba has teamed up with the Chinese Academy of Science to build a quantum computing lab. Meanwhile, various organizations (including Google) are exploring the potential of a commercial machine from D-Wave, which takes a more immediate but less powerful approach to the problem.
"This is a significant commitment from IBM," says MIT physicist and quantum computing researcher Ike Chuang, who has used a less powerful version of the quantum computer IBM plans to commercialize. "It shows a confidence that quantum computing can make money." But after years spent developing quantum technologies, IBM is also trying to prevent Google, a relative newcomer to the field, from stealing its quantum mindshare. And it's still unclear whether the claims made by these two companies will hold up. The future of quantum computing, like the quantum state itself, remains uncertain.
The promise of quantum computing is grand. At the far end of ambitious, some theoreticians believe quantum computers could crack the encryption now used to protect the world's private data. In the short-term, researchers believe quantum machines cannot only accelerate the progress of machine learning but significantly improve the development of new medications, streamline our financial markets, and even solve traffic problems.
A classical computer---a computer like the one on your desk---stores data in tiny transistors, and each of these transistors holds a single “bit” of information: a 1 or a 0. But a quantum computer changes the paradigm dramatically. Thanks to the superposition---a phenomenon exhibited by electrons, photons, and other quantum systems---a “qubit” can exists in two states at once. It stores both a 1 and a 0 simultaneously. In turn, two qubits could hold four values at once: 00, 01, 10, and 11. Taken to the theoretical extreme, you have an exponentially more powerful machine.
If that doesn't quite make sense, it shouldn't. These principles defy classical logic. Researchers first proposed the idea of a quantum computer in the 1980s and even today are only beginning to demonstrate the possibilities in reality. The paradoxical problem: If you try to examine the state of a system in superposition, it "decoheres." Rather than holding both a 1 and 0 simultaneously, it holds only a 1 or 0, like a classical machine. To build a viable quantum computer, researchers must somehow harness the probability that it will decohere into a particular state.
Google and IBM are both working to do that. IBM uses a technique where it drops superconducting circuits into a sub-zero refrigerator, a delicate process that requires extreme care. That's why the computers it will offer up to other businesses will remain in IBM facilities in New York as customers access them over the internet. In their Nature editorial, Google researchers say the company has a similar plan to offer access to its quantum processors via the cloud, building on work that grew out of the University of California, Santa Barbara.
IBM expects its customers—likely researchers at universities and big companies—to use its experimental systems in experimental ways. Like Google, the company sees its cloud efforts as a way of accelerating the rise of quantum computing. But it also wants to make a buck in the process. IBM says it has not yet decided how it will charge for its systems. It could lease them or sell them. But more likely, it will charge a regular fee for access, much like cloud computing services operate today. Regardless, the idea is to regularly upgrade these machines as the technology improves. "We don't just want to build these machines," says Jerry Chow, the former Yale researcher who helps oversee IBM’s quantum computing research. "We want to build a framework that allows people to use them."
Over the next few years, IBM hopes to expand these machines to 50 qubits, moving well beyond the 5-qubit machine it now offers over the internet (for free). At 50 qubits, the company believes, a quantum computer will handle certain tasks better than classical systems. Like Google, IBM points to chemical modeling as a possible use, which may speed drug discovery. Meanwhile, a quantum computer's ability to handle optimization and probability problems could significantly improve machine learning and financial trading, as well as solve the thorniest traffic problems. Think of it this way: when all those self-driving hits the roads, quantum computers can help ensure they all take the most efficient routes.
These improved quantum systems will still lack "error correction," meaning we won't be completely sure of every calculation they make. But researchers say that extreme power can still come amid mistakes. In Nature, Google compares this power to the recent rise of deep neural networks, complex mathematical systems that learn discrete tasks---such as image recognition and machine translation---by analyzing vast amounts of data. These systems make mistakes, and in any given situation, it's hard to completely understand how and why they work. And yet they have proven enormously effective at everything from image recognition to the most complex of games. "A lack of theoretical guarantees," the Google researchers say of quantum computing, "need not preclude success." Google and IBM are talking big. But not without uncertainty.
