Get access to Protocol
Where should we send your daily tech briefing?
The race for quantum supremacy is heating up — and also cooling down.
IBM announced Tuesday that it plans to build a 1,000-qubit quantum computer by 2023. Its machine, code-named "Condor," would far outstrip the machines available on the market today. If successful, IBM's future computer could pave the way for a revolution in computing. But it's not the only company working toward that future, and the path between the machines of today and IBM's vision is far from simple.
The company also sees this announcement as the beginning of even greater possibilities ahead. IBM said that the path it's now on should lead to IBM being able to build a million-plus qubit machine in the future. Those machines — which the company isn't putting a timeline on when it would be able to build — would potentially unlock the transformative use cases of quantum computers, from new drug discovery to super secure computers. Many see a 1 million-qubit machine as what's needed to start achieving those revolutionary uses.
But IBM still has quite a way to get to 1,000 qubits, let alone into the millions. Its current state-of-the-art machine has 65 qubits, and it plans to release a 127-qubit processor next year, and a 433-qubit device in 2022. Unlike classical bits, quantum bits are extremely fragile, and are wont to decohere, or produce errors. Current quantum computers need additional qubits to check on the calculations of other bits; for each calculation, there could be several qubits needed, which is part of the reason why so many companies in the quantum space are pushing to build such large machines.
"We think of IBM Quantum Condor as an inflection point, a milestone that marks our ability to implement error correction and scale up our devices, while simultaneously complex enough to explore potential quantum advantages — problems that we can solve more efficiently on a quantum computer than on the world's best supercomputers," IBM's vice president of quantum research, Jay Gambetta, said in a release.
IBM's research director Dar Gil sitting inside the refrigeration unit. Gil is over 6 feet tall. Photo: IBM
These future computers will also be massive. IBM is in the process of building out a refrigeration space at its Yorktown Heights research facility. It's working on a 10-foot-tall and 6-foot-wide "superfridge," internally nicknamed after the fantastic 1995 Bond film "Goldeneye," that's larger than any dilution refrigerator available on the market today. "We've designed this behemoth with a million-qubit system in mind," Gambetta said.
IBM is far from the only company working on building quantum computers that may one day be as revolutionary as the original mainframe computers were. Last year, Google claimed it had achieved a type of "quantum supremacy" — calculating something in seconds that would've taken a classical computer thousands of years to complete — and that it also had a solution in mind for how to build a million-qubit machine. Honeywell also recently claimed to have the most powerful quantum computer available on the market today, and has stated it has a pathway to far more powerful machines ahead of it over the next few years.
IBM's announcement signals the company's bold gamble to win the quantum race. What started out as a small research team just a few years ago could one day be as important to the company as its mainframe business once was. But that's a big gamble, on a strict time frame. IBM has struggled in recent years; even before the pandemic, its sales have declined year-over-year for 28 of the past 32 quarters. The company doesn't break out sales for its existing quantum services — likely a very small part of its cloud-computing services — but it's possible investors, or even new CEO Arvind Krishna, will want to see results out of the division before a 1,000-qubit machine is ready.
Mike Murphy ( @mcwm) is the director of special projects at Protocol, focusing on the industries being rapidly upended by technology and the companies disrupting incumbents. Previously, Mike was the technology editor at Quartz, where he frequently wrote on robotics, artificial intelligence, and consumer electronics.