NVIDIA is advancing its efforts in life sciences through the introduction of agentic AI, which aims to automate complex scientific workflows and enhance drug discovery processes.
At the BIO USA 2026 event in San Diego, NVIDIA's Vice President of Healthcare, Kimberly Powell, discussed how AI has evolved from merely assisting researchers with data retrieval to performing tasks autonomously. This shift towards agentic AI signifies a move from generating information to executing multi-step scientific tasks. The newly launched BioNeMo Agent Toolkit is central to this initiative, integrating various AI tools and models to facilitate capabilities across multiple domains such as biology, chemistry, and genomics.
Powell emphasized the necessity of this advanced structure in scientific research, which is often complex and multidisciplinary. The toolkit allows AI agents to effectively interact with scientific tools, enabling them to conduct tasks that would typically require human intervention, such as literature searches, hypothesis testing, and laboratory interactions. This new paradigm allows for streamlined workflows where agents can orchestrate tasks across various applications, thus enhancing efficiency in drug discovery and biomedical research.
NVIDIA's initiative also highlights the need for real-world verification in scientific applications, as digital agents must connect with laboratory instruments and experiments. The BioNeMo ecosystem, which already includes over 50 companies and research organizations, aims to create a collaborative environment that enhances the productivity of scientific work through agentic AI. As Powell noted, the transformative potential of agentic AI in science mirrors its impact on software engineering, suggesting a significant shift in how scientific research will be conducted in the future.