Nvidia CEO leads expert panels on quantum computing
Nvidia's Jensen Huang and experts set realistic expectations for quantum computing and discuss its place next to traditional high-performance computing.
Nvidia is working toward a future where quantum and traditional high-performance computing work side-by-side to solve the most difficult problems scientists face today.
On Thursday, Nvidia CEO Jensen Huang led two separate panel discussions with a total of a dozen quantum computing companies at the Nvidia GPU Technology Conference. Huang partly held the gatherings as a mea culpa for his comments in January, when he told an interviewer the industry was 15 to 30 years away from harnessing subatomic particles to process data at unimaginable speeds. His remarks sent some companies' stock into a tailspin.
"I didn't know they were public," Huang said of his earlier comments before introducing the first panel. "How could a quantum computer company be public?"
The more than two hours of discussions that followed covered various topics, including the initial beneficiaries of quantum computing, its usefulness today and how it will work with high-performance computers running GPUs and CPUs to solve problems requiring faster data processing.
"The word quantum computer is misleading because people expect that you can replace a classical computer with a quantum model. It's not like that," said Loïc Henriet, CEO of Pasqal. "They're very specialized machines you can use alongside CPUs and GPUs for specialized tasks."
Quantum computers allow us to go into corners of the universe where we have never been.
Mikhail LukinCo-founder and board member, QuEra Computing
Panelists agreed that the disciplines that would benefit from quantum computing first included chemistry, biochemistry and material science.
"Quantum computers allow us to go into corners of the universe where we have never been," said Mikhail Lukin, co-founder and board member of QuEra Computing. "I believe there is a huge potential to use the machines either existing already or being developed in the near term to really advance this scientific frontier."
Organizations are using quantum computing today to develop better refrigerants, produce hydrogen from water more efficiently, and bind peptides, which are short chains of cellular amino acids that carry out important biological functions, according to Rajeeb Hazra, president and CEO of Quantinuum.
In the future, quantum computing could help solve problems in drug discovery, global weather modeling and research on the properties of materials, said Alan Baratz, CEO of D-Wave Systems. "There are hard computational problems that are out of the reach of classical computing."
Quantum computing would also make AI more powerful by training large language models on data they wouldn't have otherwise, such as molecular configurations in a drug to determine how the human body processes it over time, Hazra said.
Panelists agreed that to get there, the first hurdle will be for researchers to build quantum processors with a million qubits, the fundamental unit of quantum data, not the few hundred that exist today. Given the current state of the quantum industry, a reasonable time for that breakthrough is 2030.
"I think the right analogy is to think of the early days of electronic computing with vacuum tubes," said Rob Schoelkopf, chief scientist and co-founder of Quantum Circuits.
Nvidia CEO's advice
Huang said he hopes the expectations for quantum computing aren't too high while it is in its infancy. When he started Nvidia, expectations were low, so people were OK with the GPU not rendering a perfect image in a computer game, the first application for the company's chips.
"They accepted it because it was a game, for crying out loud," he said. "It gave us the opportunity to scale our economics, technology and [market] footprint."
He also made sure the GPU worked alongside the PC. "We didn't replace the computer; we added to it," he said.
Pete Shadbolt, co-founder and chief scientific officer at PsiQuantum, pushed back on assuming that adding a quantum computer to a traditional system would create a more useful unit.
"There is no reason to believe that taking a small, low-performing quantum computer and plugging it into an incredibly high-performance, conventional computer is going to make things any better," Shadbolt said.
Nevertheless, Huang used a flywheel metaphor to describe how a young tech company achieves success. The flywheel starts with solving a problem better than anyone else, which results in sales that generate revenue for more R&D.
The R&D leads to a better product that drives more sales, Huang said. However, there's a caveat: "That flywheel is insanely hard to get."
Antone Gonsalves is an editor at large for Informa TechTarget, reporting on industry trends critical to enterprise tech buyers. He has worked in tech journalism for 25 years and is based in San Francisco. Have a news tip? Please drop him an email.
