Nvidia to build the U.K.'s fastest supercomputer for AI drug-hunters at GSK, AstraZeneca and more

NVIDIA, world’s 22nd fastest supercomputer, DGX SuperPOD, supercomputer
Nvidia's Cambridge-1 is expected to come online before the end of the year as the most powerful supercomputer in the U.K. and the 29th fastest in the world. GlaxoSmithKline and AstraZeneca are slated as the first to receive access. (Nvidia)

Through a new partnership with GlaxoSmithKline, AstraZeneca and the U.K.’s National Health Service, the chip maker Nvidia plans to build Great Britain’s most powerful supercomputer—and dedicate its use to artificial intelligence research in healthcare.

Dubbed Cambridge-1, the machine is designed to deliver 400 petaflops of performance, or 400 quadrillion floating-point calculations per second.

When presented with dense systems of linear equations used in AI—such as simulations of molecular models and chemical interactions among potential drug compounds—it is expected to provide 8 petaflops of supercomputing power, ranking it number 29 on the list of the world’s fastest.

Free Webinar

From Patient Adherence to Manufacturing Ease - Why Softgels Make Sense for Rx

Join Thermo Fisher Scientific’s upcoming webinar to learn why softgels offer numerous benefits for Rx drug development, including enhanced bioavailability, patient compliance and easy scale-up. Register Today.

It is slated to come online before the end of the year, with GSK and AstraZeneca among the first drugmakers to use the system. Researchers from King’s College London, Oxford Nanopore and the Guy’s and St. Thomas’ NHS Foundation Trust will also have access.

“The use of big data, supercomputing and artificial intelligence have the potential to transform research and development, from target identification through clinical research and all the way to the launch of new medicines,” said James Weatherall, AstraZeneca’s head of data science and AI, in a release.

Nvidia had announced plans last month to build an AI center of excellence in Cambridge, open to collaborations with U.K. researchers and startups. Famous for its consumer graphics cards, Nvidia's GPU chips that many use to play videogames are also well-suited in structure to power the massively parallel operations within supercomputers. 

Cambridge-1 will feature 80 networked Nvidia DGX A100 systems, with the company making an investment of £40 million, or $51.7 million U.S., in its launch.

RELATED: GlaxoSmithKline talks up its new R&D focus, keeps mum on M&A targets

“AI and machine learning are like a new microscope that will help scientists to see things that they couldn’t see otherwise,” GSK’s chief scientific officer and R&D head, Hal Barron, said in the release. “Together with GSK’s new AI lab in London, I am delighted that these advanced technologies will now be available to help the U.K.’s outstanding scientists.”

The Big Pharma’s upcoming London AI hub will also house data scientists from Nvidia, under a research partnership aimed at developing new therapeutics and vaccines using GSK’s library of genomic and biomedical data. In turn, they will provide expertise and access to Nvidia’s Clara Discovery suite of applications, aimed at computational drug discovery.

“Because of the massive size of the datasets we use for drug discovery, we need to push the boundaries of hardware and develop new machine learning software,” said Kim Branson, global head of AI and machine learning at GSK. 

Clara Discovery includes tools for analyzing medical images, genomics data and more to help build compounds and predict their responses in living tissue. In addition, using natural language processing, researchers can parse medical literature and patent documents to incorporate information on existing treatments and real-world data, according to Nvidia.

RELATED: AstraZeneca brings AI to molecule building with DeepMatter deal

Elsewhere, Nvidia and Mass General Brigham demonstrated an AI model designed to predict whether a patient with COVID-19 will eventually require supplemental oxygen therapy.

By combining analyses of chest X-ray scans and previous health records, the system aims to help triage incoming patients, as many hospitals expect to see new surges in cases during the fall and winter months.

Connecting 20 international health systems for a two-week project, the AI model was able to predict oxygen needs with 94% accuracy. This could help clinicians determine which emergency room patients may need intensive care.

Suggested Articles

A COVID-19 antibody diagnostic developed through a joint venture between Mount Sinai Health System and RenalytixAI has been authorized by the FDA.

Researchers at Northwestern University have trained an AI algorithm to automatically detect the signs of COVID-19 on a basic X-ray of the lungs.

Polyphor is developing an inhaled version of murepavadin, which targets Pseudomonas aeruginosa infections, but is currently given intravenously.