Quantum Compute Report Card: Looking Beyond Two-qubit Gates

At the start of October, IBM opened its first European quantum data center in Ehningen, Germany, which brought the company’s available quantum compute resources to the continent's academics, researchers, and other organizations. It marked not only an expansion of IBM’s reach in the development of quantum computers, but the ubiquitous interest in taking advantage of the technology -- even though it is still years away from commercial use.

Expectations heaped on quantum computers can range from optimistic hopes to fear of its security-smashing capabilities if put into the hands of bad actors. The multiyear, industrywide effort to further the development of the technology might not have the flash of AI-related deployments, but the era of quantum computing and post-quantum cryptography may have deep, lasting ramifications for cybersecurity and beyond.

“Quantum computers from a user perspective didn’t exist until, like, 2016 when IBM put one upon the cloud,” says Scott Crowder, vice president of IBM quantum adoption and business development. “This is not science fiction. This is a real computer, and you can go use it yourself.”

IBM holds a significant role in the development of quantum computing, but it is far from the only stakeholder. This latest, two-part “report card” on the technology’s current progress will continue with a follow-up story on Tuesday with more industry watchers and players in the development of quantum computing.

Related:Are You Ready for the Attack of the Copper Thieves?

How Quantum Is Shaping Up

In 2023, Crowder says, the next major milestone emerged once a quantum computer with a software stack ran a computation with accuracy that could not be simulated classically.

“So, a quantum computer is better at being a quantum computer than a classical simulator ever will be,” he says. The next check, Crowder says, is to perform a computation that has individual use that cannot be done any other way. “To me that’s some variation of quantum advantage,” he says, which is a step the industry is working towards.

For its part, IBM’s roadmap includes showing how increasingly complex programs can be run as the technology develops. “The next time we do our green checking mark activity, we’re going to put the green checking mark next to that -- we’ve delivered systems that our users can run 5,000 two-qubit gate operations and get a quality, accurate result,” Crowder says. “That’s bigger than you can possibly classically simulate.”

He expects somewhat incremental growth in those two-qubit gate operations, year-by-year, that can be run on the way to a profound leap. “That big jump to 100 million [two-qubit gate operations] is when we introduce error correction, fault tolerance … that allows you to basically make a discontinuity and be able to run much, much, much longer programs.”

Related:New Data Center Developments: December 2024

Crowder says determining the next major checkpoint for quantum advantage, where users should be able to run a computational problem that has scientific benefit with a result that could not be attained any other way, is a hard question to answer. “We’re in this weird situation now where we’ve built computers that are bigger than anyone can possibly simulate, but they’re smaller than what I can prove to you on a blackboard will be better,” he says. “I need to actually run it on the computer to see if it’s better. So, I can’t prove it to you until I prove it to you.”

Keys to the Quantum Kingdom

In order to achieve those next checkpoints and draw closer to delivery of quantum computing’s promises, Crowder says three elements are needed.

“One of them is the underlying hardware technology,” he says. “We think we know all the pieces we need to build. We think we’ve demonstrated all the pieces from a research point of view. Now we only need to, from an engineering and development point of view, put those pieces together and still hold the value.”

Related:Clearing the Clouds Around the Shared Responsibility Model

Last year, Crowder says, IBM put out a lower error rate, gate architecture that continues to drive down their gate errors.

The second element, he says, is that many people underappreciate the role software -- and coming up with software that is efficient on classical resources -- plays. “When you talk about error correction and fault tolerance, people hyper-focus on surface code, which probably isn’t the right approach,” Crowder says. That might not factor in how to efficiently scale error coding so that software can run on regular computers efficiently enough that it will not take multiple years just to decode what happened on a quantum computer. “That’s actually a fairly large challenge.”

The third piece needed to reach the major checkpoint, he says, is algorithmic development, which might call for herculean effort in some regards. “Understanding like the algorithmic approach, understanding the problem representation and how to map your problem to the underlying compute is where this gets unlocked,” Crowder says, “and if anything, I think that’s probably my biggest worry is probably … accelerating that one. It’s a talent problem and it’s a focus problem.”

The next part of this quantum report card, coming tomorrow, will dive further into Crowder’s insights as well as open the door to other players and stakeholders in the growth and future of quantum computing.