Nvidia Joins Ongoing Race in Quantum-Computing Cloud Services

Jane Lanhee Lee, Bloomberg News

Wires connected to a supercomputer built with Nvidia Corp. AI processors. , Photographer: Michelle Bruzzese/Bloomberg Wire

(Bloomberg) -- Nvidia Corp. launched a cloud service for researchers to test out their quantum-computing software, seeking to profit from a field that’s winning funding around the world despite yielding few groundbreaking applications so far.

The Nvidia Quantum Cloud will first comprise a data center stacked with AI chips and systems that together simulate a quantum computer, according to Tim Costa, director of high performance computing and quantum computing at the company. Unlike other cloud services, Nvidia’s has no quantum computer attached to it at the moment, but it will provide access to third-party quantum computers in the future, Costa told Bloomberg ahead of the GPU Technology Conference Monday.

The AI chipmaker’s cloud is years late to an already crowded arena. Microsoft Corp. started Azure Quantum in 2019, with Amazon Web Services launching Braket that same year. IBM Corp.-backed startup Strangeworks Inc. was founded in 2017 and has dozens of quantum computers on its cloud. But Alphabet Inc.’s Google Cloud, Microsoft Azure and Oracle Cloud Infrastructure will integrate Nvidia’s quantum simulator into their offerings, Costa said.

Read more: Quantum Computing Is Finally Here. But What Is It?: QuickTake

The world’s most powerful countries and corporations are pledging billions of dollars to quantum computing research, although there are few known instances where a quantum computer has run practical applications. The global zeal stems from the promise of computation speeds that are hundreds of millions times faster than supercomputers today, along with fear of the consequences — both military and commercial — for those who miss out.

For most of the 2010s, Nvidia was largely a bystander in that push. But it’s now a key player in quantum research after its graphics processing units, designed to speed up video games, paid dividends thanks to their ability to execute a certain type of computation quickly.

A similar version of these calculations is used in quantum computer simulations, making Nvidia AI chips such as the H100 Tensor Core GPUs an important component of quantum research. The GPU is now also powerful enough to run classical algorithms once too big for computers, such as those used for drug and materials discovery and other scientific domains, said Costa.

Separately, Nvidia also said that its chips will be used in the publicly-funded ABCI-Q supercomputer that Fujitsu Ltd. is building for the quantum computing research arm of Japan’s National Institute of Advanced Industrial Science and Technology. The supercomputer will use more than 2,000 Nvidia H100 Tensor Core GPUs and will be deployed early next year.

The company said that supercomputers at Denmark’s Novo Nordisk Foundation and Australia’s Pawsey Supercomputing Research Centre, both dedicated to research into quantum computing and other fields of cutting-edge technology, will also use Nvidia hardware.

(Updates to clarify supercomputer usage in final paragraph)