NVIDIA video: Lila Sciences' AI agents and automated labs for drug discovery
NVIDIA's official channel released a video detailing how biotech firm Lila Sciences uses NVIDIA Nemotron and Agent Toolkit to build experimental plans, filters candidates via millions of virtual experiments using BioNeMo, and validates them with robotic instruments in its AI Science Factory. The resource illustrates the specific workflow connecting AI agents to autonomous labs.
NVIDIA published a video on its official YouTube channel on October 21, 2026, introducing how biotechnology company Lila Sciences combines artificial intelligence agents with automated laboratories to accelerate drug discovery. This resource offers reference value for professional readers interested in the concrete implementation of "AI for Science."
According to the video description, Lila Sciences is exploring a therapy designed to help a patient’s own immune cells fight B-cell lymphoma. The workflow consists of two main stages: first, building an experimental plan using NVIDIA Nemotron and NVIDIA Agent Toolkit, while models built with NVIDIA BioNeMo run millions of virtual experiments to identify promising candidate molecules. Subsequently, these candidates are sent to Lila’s "AI Science Factory," where robotic instruments manufacture and test them.
The core of this case lies in the data feedback loop: results from physical lab tests return to inform the next round of designs and experiments. Although this entry is based solely on the publisher's descriptive information and does not include full interview dialogue or independent reproduction data, it clearly outlines the typical architecture of current AI-driven drug development involving "computational screening + automated wet-lab validation."