Virtual Lab designs 92 COVID-19 nanobodies with promising results
Updated
Updated · The Stanford Daily · May 6
Virtual Lab designs 92 COVID-19 nanobodies with promising results
10 articles · Updated · The Stanford Daily · May 6
Stanford researchers said two candidates bound effectively to both newer variants and the original virus after Biohub testing, with the core AI planning completed in one to two hours.
The multi-agent system, led by James Zou and Kyle Swanson, chose nanobodies over conventional antibodies and generated designs in days rather than the weeks or months typical for human researchers.
Researchers said the system still needs wet-lab scientists to judge practical constraints and run experiments, while related work has expanded into a larger Virtual Biotech using thousands of agents.
AI can now discover drugs faster than humans. Is the era of the human scientist coming to an end?
When an AI research team can be hacked, how do we protect scientific discovery from digital sabotage?
If an AI invents a blockbuster drug, who legally owns the discovery and the multi-billion dollar profits?
92 AI-Designed Nanobodies: How Virtual Labs Are Revolutionizing COVID-19 Therapeutics and Biomedical Discovery
Overview
The Virtual Lab is a groundbreaking AI-human research collaboration that emerged in 2026, transforming biomedical discovery. By assigning autonomous AI agents to specialized roles—such as principal investigator, domain specialists, and scientific critic—the Virtual Lab minimizes human involvement and accelerates scientific breakthroughs. These AI agents drive most of the intellectual work, working together to solve complex problems. A major achievement was the creation of a novel computational pipeline for designing nanobody binders against evolving SARS-CoV-2 variants, demonstrating the Virtual Lab’s ability to rapidly generate innovative solutions to urgent health challenges.