Nvidia Corp. is teaming up with biotechnology giant Genentech Inc. on a planned multiyear collaboration that will see the two companies accelerate research into drug discovery by leveraging the most advanced artificial intelligence capabilities, including generative AI.

The company said it’s hoping to enhance Genentech’s advanced AI research programs by transforming its generative AI models and algorithms into a “next-generation AI platform” that can expedite the discovery of novel therapies and medicines.

Genentech has already developed its own proprietary machine learning algorithms and models, but it plans to tap Nvidia’s expertise to significantly improve on them. The idea is to use Nvidia’s DGX Cloud, which offers a training-as-a-service platform based on the chipmaker’s dedicated AI hardware, including its graphics processing units, and software such as Nvidia BioNemo.

BioNeMo, now generally available as a training service, is a new, domain-specific platform that simplifies, accelerates and scales up generative AI applications for computational drug discovery. It allows researchers to pretrain or fine-tune state-of-the-art models on DGX Cloud.

As part of the collaboration, Nvidia also plans to share its computing expertise. It said it will work closely with Genentech’s teams of computational scientists with the goal of optimizing and scaling up its AI models, with one of the side benefits that it may also be able to improve its own platforms.

According to Genentech, the process of drug discovery and development is an extremely lengthy and complicated one, not to mention very uncertain. For scientists, it remains extremely difficult to predict targets for novel medicines, the company said. It’s also incredibly difficult to develop molecules that have potential as therapeutics.

However, the company believes AI can play a key role in making all of this much easier. It says AI will help drug discovery to become more predictable and cost-effective, and boost the success rate of its research and development programs over the long term.

Aviv Regev, executive vice president and head of research and early development at Genentech, said the initiative will help his company to unlock new scientific discoveries with “incredible speed” and generate insights at scale. “Bringing science and technology together has always been a foundation of biomedical breakthroughs at Genentech,” Regev said. “We are thrilled to join forces with Nvidia to further optimize our drug discovery and development to deliver treatments that transform people’s lives.”

Genentech’s AI teams are busy developing AI and machine learning foundational models across numerous research areas, including diverse therapeutic modalities. With this, their goal is to glean insights for target and drug discovery, while answering fundamental questions regarding human biology and disease.

Another goal of Genentech is to accelerate its “lab in a loop” initiative, which is focused on using experimental data to feed computational models that can uncover patterns and make new, experimentally-testable predictions. According to the company, its scientists will now be able to assess these predictions much more rapidly using Nvidia’s platforms. The results can then be fed back into its underlying computational models to improve them, enabling the development of more effective therapies.

The collaboration will leverage Genentech’s proprietary data as well as publicly available datasets. Nvidia itself will not have access to Genentech’s data, unless it’s specifically granted permission for use in a particular project.

Photo: Genentech