>FASTA_2

NVIDIA's biotech tools, NeoPlants GM House Plants, Helix Nano's Immune-Computer Interface

Digital biology, by NVIDIA

NVIDIA is a company that designs GPUs (chips) and software platforms for data science and video games, and has multiplied its value by 42,000% since going public. Its CEO, Jensen Huang, founded it in 1993 with the core belief that 3D graphics would become a fundamental technology for the video game industry.

Almost three decades later, the necessary elements have come together—primarily the vast supply of biological data resulting from the reduction in biotechnology costs and more advanced artificial intelligence models—to create what could be the best market NVIDIA serves: digital biology.

Today, on its CUDA-X™ platform, you can find services like Parabricks®, which analyzes genomes en masse up to 100x faster than platforms that use only CPU technology. This means analyzing a complete human genome (with 30x coverage) in 10 minutes instead of 24 hours while obtaining the same quality of results. Its users include companies like Agilent Biotechnologies.

Continuing with the topic of last week's “BioChatGPT,” BioNeMo is a generative AI platform that offers workflows and models like pre-trained ESM-Fold to predict protein structures, DNABERT to understand the function of non-coding genes, or scBERT that uses single-cell information to predict the effects of a knockout. Clients of these include Amgen and AstraZeneca.

Example of ESM-Fold running through BioNeMo

The design capability of its chips allows NVIDIA to offer the fastest and most integrated platform to run these models. Furthermore, it has found great allies in companies like Amazon for storing this data. Additionally, models like ESM have been developed by companies like Meta. There is definitely a fascinating case for these and many other big techs joining forces to penetrate the biotech market.

NeoP1, by NeoPlants