Nvidia says its new NVIDIA Quantum Cloud is designed to help researchers in biopharma and other branches of science advance quantum computing and algorithm research. For the first time, according to Nvidia, users can use the cloud to build and test new quantum algorithms and applications—including simulators and tools for hybrid quantum-classical computer programming.
Despite its name, Quantum Cloud does not include a link to a quantum computer, though Nvidia plans to let users access quantum computers from outside companies in the future.
Those tools will be available to users as “microservices” through top-tier public cloud platforms such as Amazon Web Services (AWS), Google Cloud, Microsoft Azure, and Oracle Cloud Infrastructure—as well as through the offerings of leading quantum computing companies, such as IQM Quantum Computers, OQC, ORCA Computing, qBraid, and Quantinuum.
“We are potentially the largest quantum computing company in the world that doesn’t build a quantum computer,” Nvidia founder and CEO Jensen Huang declared during a question-and-answer session with reporters at the conference, held at the San Jose Convention Center. “The reason why we do that is because we believe in quantum computing.
We don’t think that it’s necessary for us to build yet another one. Quantum computing is more than just a quantum computer.”
NVIDIA Quantum Cloud is based on the company’s open-source CUDA-Q™ quantum computing platform. According to Nvidia, CUDA-Q is used by three of every four companies that deploy quantum processing units (QPUs).
Quantum Cloud will offer users access to Nvdia’s cuQuantum software, which is designed to simulate QPUs on graphics processing unit (GPU)-based computer hardware. Nvidia reasons that a simulator will offer the best opportunity for developers to explore how quantum technology will impact scientific exploration because QPUs have yet to achieve the power and stability that is required to test quantum algorithms and applications by themselves.
Integrating third-party software
NVIDIA Quantum Cloud integrates the software of several third-party developers. Classiq Technologies, a startup developer with offices in Tel Aviv, Boston, and Tokyo, has integrated its quantum computing software offering with CUDA-Q. The integration allows quantum researchers to generate large, sophisticated quantum programs, as well as to deeply analyze and execute quantum circuits.
“It takes less in resources to simulate a quantum system on a quantum computer than it does on a classical computer. We’re getting to a point where you cannot simulate large molecules on a quantum computer, regardless of the resources. It’s just too computationally intense,” Erik Garcell, PhD, head of technical marketing with Classiq, told GEN from the Expo Hall of the NVIDIA GTC 2024 conference.
Because research is central to the work of drug developers, he said, “what they can hopefully get to is allowing the simulation of larger molecules, like how a protein might interact with the human enzyme.”
“If you could get to a point where you could simulate this, you can accelerate the speed of drug discovery. You could democratize drug discovery, allowing more companies to develop new drugs, hopefully,” Garcell added.
Nvidia offered two other examples of integrations of third-party software into Quantum Cloud for applications in life sciences and other research.
QC Ware, a provider of software and services based in Palo Alto, CA, has agreed to integrate its Prometheum molecular discovery platform with Quantum Cloud, with the aim of accelerating drug discovery. The partnership is intended to provide AI platforms with the ability to generate highly accurate training data on large, complex molecules faster than ever before, and thus reshape in silico molecular simulation.
“Promethium is using NVIDIA Quantum Cloud to rapidly generate training data on large, complex molecules, enabling better machine learning models that will help pharmaceutical companies find drug candidates more quickly,” stated Tim Costa, Nvidia’s director of HPC and quantum.
The partnership also aims to help molecular AI platforms train better ML models, ultimately helping pharmaceutical companies find quality drug candidates more quickly, QC Ware said.
Also integrated with Quantum Cloud is the Generative Quantum Eigensolver (GQE), a novel method for applying classical generative models for quantum simulation developed by researchers at the University of Toronto in collaboration with Nvidia. First disclosed in a preprint posted in January, GQE uses large language models (LLMs) to enable a quantum computer to support molecular research by finding the ground-state energy of a molecule more quickly.
Powering supercomputers
Nvidia also announced several collaborations that are intended to advance the quantum-based supercomputing efforts of a pair of international partners.
The supercomputer features more than 2,000 NVIDIA H100 Tensor Core GPUs in 500+ nodes interconnected by NVIDIA Quantum-2 InfiniBand, a fully offloadable, in-network computing platform that Nvidia says is the only one of its kind in the world. ABCI-Q is expected to be deployed early next year and is designed for integration with future quantum hardware.
ABCI-Q is designed to serve a platform for advancing quantum circuit simulation and quantum machine learning, the building of classical-quantum hybrid systems, and the development of new algorithms inspired by quantum technology. The supercomputer is being built by Fujitsu at Japan’s Global Research and Development Center for Business by Quantum-AI Technology (G-QuAT) National Institute of Advanced Industrial Science and Technology (AIST) ABCI supercomputing center.
In Denmark, the Novo Nordisk Foundation and the Export and Investment Fund of Denmark (EIFO) are leading a national collaboration with Nvidia that is designed to establish a national center for AI innovation that will house one of the world’s most powerful AI supercomputers.
The Danish Centre for AI Innovation is expected to be ready for pilot projects before the end of the year.
$100M+ commitment
The Novo Nordisk Foundation, which controls Novo Nordisk, has committed approximately DKK 600 million ($87.4 million) towards the initial costs of the center, while EIFO has contributed DKK 100 million (about $14.6 million).
Denmark aims to accelerate research and innovation in life sciences as well as healthcare and sustainable development, through the development of the supercomputer, to be called Gefion. Denmark’s new supercomputer will consist of a large-scale NVIDIA DGX SuperPOD, which will be powered by NVIDIA H100 Tensor Core GPUs and interconnected using NVIDIA Quantum-2 InfiniBand networking.
“Drug discovery, disease diagnosis, and treatment, as well as complicated life science challenges, are examples of areas where extreme AI computing power can enable the positive transformation of our society,” Mads Krogsgaard Thomsen, CEO of the Novo Nordisk Foundation, said in a statement. “The collaboration with NVIDIA and the resulting national AI innovation center can help Denmark’s brilliant researchers and innovators rise to the next level.”
At present, no GPU-accelerated supercomputers operate in Denmark—a factor pinpointed as an obstacle to AI research by researchers, universities, and government officials, according to the foundation. Eviden, an Atos Group company specializing in advanced computing, will deliver, install, and configure the supercomputer later this year, and provide support with the start-up phase.
Nvidia counts more than 160 partners in its quantum computing ecosystem.
Nvidia unveiled its microservice approach to delivering technology tools at the start of GTC 2024 when it unveiled a an expansion of its NVIDIA BioNeMo™ generative AI platform for drug discovery with foundation models capable of analyzing DNA sequences, predicting how proteins will change shape in response to a drug molecule, and determining a cell’s function based on its RNA.
GEN’s’ attendance at NVIDIA GTC 2024 was funded by Nvidia.
