10 quantum computing milestones of 2024

8. RIKEN and NTT launch world first general-purpose optical quantum computer

RIKEN and NTT launched the world’s first general-purpose optical quantum computer in November. The system operates at nearly room temperature and processes at speeds up to several hundred terahertz.

The computer uses continuous-variable analog design with time-division multiplexing, where computations occur through quantum teleportation. It can handle approximately 100 continuous quantity inputs. The system is accessible through a cloud service and is designed for materials science, chemistry, and AI applications.

One major benefit of this approach to quantum computing is that it operates at room temperature. By comparison, superconducting qubits and trapped ion cubits operate at temperatures close to absolute zero. According to Riken, the optical quantum computer is also faster than other quantum computing platforms.

9. D-Wave speeds up processing 25,000 times

And in yet another kind of quantum computing — the analog kind — D-Wave released the benchmark results for its latest 4,400-plus qubit Advantage2 processor in November, with big performance gains. The system solves materials science problems 25,000 times faster than its previous version, the company claims.

The new processor doubles qubit coherence time and has improved qubit connectivity, enabling solutions to larger problems.

The system delivered five times better solutions for high-precision applications and outperformed the previous version in 99% of satisfiability problem tests, the company said. It’s available now through D-Wave’s quantum cloud service.

D-Wave’s qubits can’t be directly compared to those of other companies, however, because it’s not a general-purpose computer but a narrow-purpose analog one.

10. Google uses AI to improve error correction

We can expect to see even more progress on error correction next year. One reason? Artificial intelligence.

In late November, Google announced an AI system that spots errors in quantum computers. This uses the same transformer technology that’s behind large language models and other types of generative AI, but trained specifically on quantum processes.

The system is too slow to be actually used in real-time to spot errors, Google says, but, as AI technology improves, it could point to an additional way to reduce error rates.

The road ahead

So, when will we have working quantum computers?

Analysts at the Boston Consulting Group say they’ve been disappointed in the lack of recent progress of actual enterprise use cases.

“Quantum computing today provides no tangible advantage over classical computing in either commercial or scientific applications,” Boston Consulting Group researchers wrote in a July report. “Though experts agree that there are clear scientific and commercial problems for which quantum solutions will one day far surpass the classical alternative, the newer technology has yet to demonstrate this advantage at scale.”

According to the firm, our current era, that of “noisy intermediate-scale quantum” will last until 2030.

At this stage, quantum computers have too few qubits and too many errors to be practically useful, but they can be used to test algorithms and do other preliminary research.

Broad quantum advantage — where quantum computers can be used to solve problems that traditional computers can’t handle — will last from approximately 2030 to 2040, according to the Boston Computing Group. At that time, the research firm predicts that end users will see between $80 and $170 billion in total annual value creations.

In 2040, according to the research firm, we’ll enter the era of full-scale fault tolerance, and end users will see between $450 and $850 billion of value creation per year. This is when errors are reduced to the point that quantum computers can be scaled up to handle pretty much anything thrown at them.

Some quantum computing companies expect to get there even faster. For example, IBM’s quantum roadmap has the company delivering a fully error-corrected system in 2029. It expects to have a quantum computer with over 2,000 error-corrected qubits — or logical qubits — after 2033.

It takes many physical qubits to create one error-corrected logical qubit, so IBM expects to have a 100,000-qubit computer by then, capable of running 1 billion gates.

IBM is the company to watch, says Gartner’s Dekate. “It’s very clear that IBM is leading,” he says. “They’re innovating at every layer of the stack. But it is early days.” Achieving workable quantum computers is a marathon, he says. “And we’re in the first five minutes.”

Whoever is ahead today might not necessarily win the race, he says. “Enterprises should not fall into the trap of trying to figure out who’s ahead and who’s behind — that will lead enterprises to make bad decisions, because the leaderboard will change quite dramatically.”

He suggests that companies looking to take the lead in finding ways to take advantage of quantum computing should partner with multiple vendors and explore different pathways. And, instead of focusing on finding short-term investment returns, they should instead think about skills development, building capabilities, investigating use cases, and preparing for disruptions.