Nvidia releases open-source BioNeMo Agent Toolkit for drug discovery and life sciences
Drug researchers get a new free toolset. Nvidia $NVDA released its BioNeMo Agent Toolkit, open-source software that lets AI agents and lab scientists gather evidence, run computational experiments and flag next steps to speed discovery. Available now on GitHub and Nvidia's develo
Nvidia moved deeper into the lab on June 23, releasing the BioNeMo Agent Toolkit, open-source software aimed squarely at drug discovery and life sciences research. For scientists and biotech teams, the pitch is concrete: the toolkit packages more than a decade of Nvidia's life sciences libraries, tools and open models into a single resource that AI agents can use to gather evidence, reason across findings, run computational experiments and recommend the next best step. Rather than replacing researchers, the system is designed to work alongside them, handling the heavy computational lifting while humans steer the science. The release lands in the middle of a broader shift toward agentic AI in healthcare, where models are being asked not just to answer questions but to chain together multi-step research tasks. Nvidia $NVDA says the toolkit and its skills are available now through the company's developer resources page and on GitHub, lowering the barrier for smaller labs that cannot build this infrastructure themselves. On the same day, Databricks introduced Genesis Workbench, an open, modular blueprint that integrates Nvidia tools including BioNeMo and Parabricks into one environment for end-to-end drug discovery. Together the two launches point to a tighter stack between chipmakers, cloud data platforms and the pharmaceutical researchers who ultimately run the experiments. For an industry where a single drug program can run for years, the value proposition is speed: getting from raw biological data to a testable hypothesis with fewer manual steps.