Eli Lilly, Nvidia to construct supercomputer, AI factory for drug discovery

Eli Lilly and Nvidia are partnering to construct what they call the pharmaceutical industry’s “strongest” supercomputer and so-called AI factory to assist speed up drug discovery and development across the sector, the businesses announced Tuesday. 

It’s the most recent stride by Nvidia and the pharmaceutical industry to harness AI to assist shorten the time it takes to bring cures to patients, while reducing costs at every stage of drug discovery and development. The method typically takes about 10 years on average from dosing the primary human with a drug to its launch in the marketplace, said Diogo Rau, Eli Lilly’s chief information and digital officer, in an interview. 

Eli Lilly expects to finish the buildout of the supercomputer and AI factory in December. They may go surfing in January. But the brand new tools likely won’t yield significant returns for the corporate’s business and that of some other drugmaker until the top of the last decade.

“The things that we’re talking about discovering with this type of power that now we have at once, we’re really going to see those advantages in 2030,” Rau said. 

The industry’s efforts to make use of AI to bring medicines to people faster are still within the early stages. There are not any drugs in the marketplace designed using AI, but progress is obvious within the variety of AI-discovered drugs entering clinical trials, recent AI-focused investments and partnerships amongst drugmakers.

Eli Lilly will own and operate the supercomputer, which shall be powered by greater than 1,000 Blackwell Ultra GPUs – a more recent family of chips from Nvidia – connected on a unified, high-speed network. The supercomputer will power the AI factory, a specialized computing infrastructure that can develop, train and deploy AI models at scale for drug discovery and development.

The supercomputer “is absolutely a novel scientific instrument. It’s like an infinite microscope for biologists,” said Eli Lilly’s Chief AI Officer Thomas Fuchs. “It really allows us to do things we couldn’t do before at that big scale. 

Scientists will have the ability to coach AI models on tens of millions of experiments to check potential medicines, “dramatically expanding the scope and class” of drug discovery, in keeping with a release from Eli Lilly. 

While finding recent drugs is not the only focus of the brand new tools, it’s “where the massive opportunity is,” said Rau.

“We’re hopeful that we’ll have the ability to find recent molecules that we never would have with humans alone,” he said. 

Several AI models shall be available on Lilly TuneLab, an AI and machine learning platform that enables biotech firms to access drug discovery models that Eli Lilly has trained on years of its proprietary research. That data is price $1 billion.

Eli Lilly launched that platform in September as a solution to expand access to drug discovery tools across the sector. 

“It’s really powerful to have the ability to present that extra start line to those startups that, you already know, otherwise could take a few years burning their capital to get to that time,” said Kimberly Powell, Nvidia’s vp of health care, adding that the corporate is “delighted to participate” in that effort. 

In exchange for access to the AI models, biotech firms are expected to contribute a few of their very own research and data to assist train them, Rau noted. The TuneLab platform employs so-called federated learning, which suggests that firms can reap the benefits of Lilly’s AI models without either side directly sharing data.

Eli Lilly also plans to make use of the supercomputer to shorten drug development and help get treatments to people faster. 

The corporate said recent scientific AI agents can support researchers, and advanced medical imaging can provide scientists a clearer view of how diseases progress and help them develop recent biomarkers — a measurable sign of a biological process or condition — for personalized care. 

“We’d actually prefer to deliver on that promise of precision medicine,” Powell said. “Without an AI infrastructure and foundation, we’ll never get there, right? So we’re doing all the essential constructing, and now we’re seeing this true lift off, and Lilly is a precise example of that.”

Precision medicine is an approach that tailors disease prevention and treatment in keeping with differences in an individual’s genes, environments, and lifestyles.