When will quantum computing be available? It depends

The current state of quantum computing

There's no widely accepted count of the installed base of quantum computers, but various estimates put the population in the low hundreds. D-Wave Quantum, IBM, IonQ and Quantinuum are among the vendors that have deployed quantum machines.

Such offerings are available today through cloud platforms and on-premises installations. Martina Gschwendtner, a consultant in the Munich office of McKinsey & Co., said she expects industry customers to access quantum machines through the cloud, while on-premises deployments will be more typical for research and academic institutions.

Scott Crowder, vice president of quantum adoption at IBM, said the company offers a shared-access model, noting that most of its quantum customers use its machines in the cloud. Since 2016, IBM has rolled out more than 75 quantum computers for cloud access. The company frequently updates those machines, so about 10 to 15 computers are available for use at any given time, Crowder said.

On the on-premises side, the company has installed "a couple handfuls" of dedicated systems that are available to specific customers and their ecosystem of users. Those customers include the Cleveland Clinic, the Platform for Digital and Quantum Innovation of Quebec, and Japan's Riken Center for Computational Science.

Michael Biercuk, founder and CEO of Q-Ctrl, a quantum infrastructure software company, pointed to a variation in the on-premises model: a "component-based systems" approach that lets customers assemble their own quantum machines. The company's software automates the process of building a functional quantum computer from a bare-metal chip. Q-Ctrl partners with Rigetti Computing and QuantWave, which manufacture quantum processor chips, to offer component-based quantum computing.

Regardless of deployment method, quantum computing customers thus far have focused on understanding the essentials of the technology and building expertise, noted Sam Lucero, formerly chief analyst for quantum computing at Informa TechTarget's Omdia market research group.

That experimentation phase is edging toward operational use among some early adopters.

"These adopters could probably do the same job, maybe even better, on classical computers, but they want to really jump in with both feet and start using [quantum technology] on problems of interest now," Lucero said. "I wouldn't characterize this as a sharp binary of experimentation-operation. I would characterize it more as a spectrum-based journey, and they are creeping along this spectrum further and further."

Biercuk, meanwhile, pointed to the arrival of generative AI as an analog to what's happening now in the quantum adoption cycle.

"With ChatGPT, many organizations were caught completely flatfooted, while others had already been experimenting with various forms of machine learning, reinforcement learning and AI," he said. "Those who prepared for ChatGPT had in-house skills, comfort with a new set of emerging tools and could move much faster than their peers who didn't know what it was."

Early quantum adopters are moving aggressively to gain a similar edge, Biercuk noted.

When will quantum computing become available?

Quantum computing has been in the early availability/technical feasibility stage for several years. The earliest quantum computers were based on an annealing technique and geared toward optimization problems. D-Wave sold such a machine to Lockheed Martin in 2011. Gate-based quantum computers, considered more versatile, followed. Crowder noted that IBM made its quantum technology available in the cloud in 2016.

"The first stage is quantum computers aren't just sci-fi -- they really exist," Crowder said. "You can run programs on them. People can access them."

Quantum supremacy

The next rung up the quantum ladder is where things start to get tricky. A few quantum computer providers have claimed quantum supremacy. That's the point at which a quantum computer can perform tasks that a classical computer can't. In March, D-Wave said its annealing quantum computer bested a classical supercomputer in solving a magnetic materials simulation problem. Also in March, a group of physicists in China said their Zuchongzhi 3.0 prototype quantum computer achieved quantum supremacy.

The first stage is quantum computers aren't just sci-fi -- they really exist.

However, assertions of quantum supremacy -- or any form of technical breakthrough, for that matter -- are often disputed. For instance, vendor claims of achieving a certain number of logical qubits, which is important for the accuracy of computational results, invite scrutiny.

"A lot of times, it gets really murky," Lucero said. "There will be an announcement, 'We've demonstrated X number of logical qubits.' You wait a few days, you wait a week, and the scientific community weighs in, and it turns out there are more of these caveats and nuances."

Quantum advantage

So-called quantum advantage is considered the next step; here, the definitions begin to take on different flavors.

"Looking at the question of quantum advantage, that seems to be sort of fragmenting," said Carl Dukatz, global lead of the quantum program at Accenture. "It can be somewhat of a moving target."

Lucero said he sees two levels of quantum advantage: quantum economic advantage and quantum computational advantage. Quantum economic advantage is the ability to gain some cost, accuracy or energy-efficiency edge over classical computing. This level might also entail a slight speed advantage, but not something classical computers can't achieve, he added.

Whether a machine has quantum economic advantage is almost entirely subjective, which makes the timeline ambiguous. Adopters ultimately decide whether they see a benefit relative to the computing platforms they normally use.

"It's very much in the eye of the beholder," Lucero said.

To that end, Accenture and MIT provide an online calculator that helps users gauge whether it makes economic sense to solve a particular problem on a quantum computer.

"It's important from the business perspective to understand when [quantum technology] will be cheaper or provide a comparable outcome," Dukatz said.

The second level is quantum computational advantage. This form of advantage occurs when a quantum computer performs something computationally that could not be done on a classical computer on any practical timeline, Lucero noted.

Quantum computational advantage sounds somewhat like quantum supremacy. But Lucero links fault tolerance to the arrival of computational advantage. Fault tolerance, among the main challenges in quantum computing, addresses the key quantum computing problem of high error rates.

"Computational advantage is really going to be large-scale, fault-tolerant quantum computers being able to run those circuits with only one error in a trillion operations, one error in several billion operations," he said.

That degree of fault tolerance is generally considered to be several years away. In the meantime, quantum advantage by other reckonings could be closer at hand.

At IBM, Crowder described quantum advantage as "running a computation using a quantum algorithm on a quantum computer that is better in some way than anything I can do classically."

Crowder said IBM or one of its partners will demonstrate this form of quantum advantage this year or next.

Quantum practicality

Q-Ctrl's Biercuk, meanwhile, sees another milestone -- which might be described as quantum practicality -- coming fairly soon. The availability of practical quantum computing will happen when a business or researcher actively selects a quantum computer over an alternative computing platform to solve a problem, he said.

"This is not a computer science theoretic argument," Biercuk said. "It's simply about what you would choose."

Biercuk said his company has investigated when quantum machines will be big enough, as measured in qubit counts, to tackle problems that are too hard for conventional computers.

"The surprise to us is that it looks like the machines only have to be a little bit bigger than they are today," he said. "That is, 300 to 500 qubits."

The sweet spot appears to be 300 to 500 qubits, Biercuk said, noting that IBM already has 156-qubit machines commercially available.

"This is not a far-off horizon," he said. "This looks to us like a two-to-three-year time horizon, where a quantum computer can do things that are very hard for a classical computer and are relevant."

A quantum-relevant machine would deliver high economic or strategic value to certain types of users or solve problems historically underserved by conventional computing, according to Biercuk.

Gschwendtner views quantum advantage, or "Q Day" as she calls it, as having economic and computational components that can apply to real-world use cases and problems. She estimated that the realization of those goals will happen around 2030, plus or minus a year or two.

Advancements in quantum computing

The pace of technology development will influence an earlier or later arrival date for quantum advantage.

As it happens, those developments appear to be accelerating. The top hyperscalers -- AWS, Google and Microsoft -- have all announced quantum processors in recent months. AWS was the latest to enter the market, launching its Ocelot quantum chip in March.

Technology development, in general, is moving ahead with no indication that vendors will fall significantly behind their envisioned roadmaps, noted Henning Soller, a partner and quantum research leader at McKinsey.

In addition, the funding environment for quantum computing appears to be steady.

"There's no sign of a quantum winter in terms of the investment volumes," he said.

Beyond funding, there's a long list of variables that could influence the availability of quantum advantage. Among those are improvements in quantum gate fidelity (the higher the fidelity, the lower the error rate), qubit connectivity and the ability to control qubits at scale, Gschwendtner said.

Manufacturing processes are an additional factor, she noted. Can an emerging quantum technology use an existing manufacturing technique, or will it require a new fabrication facility? Another issue: the quantum talent gap. Gschwendtner said universities are launching programs to cultivate quantum personnel, noting that enough quantum-ready employees should be available by 2030.

John Moore is a writer for Informa TechTarget covering the CIO role, economic trends and the IT services industry.