37,000-Agent AI Virtual Biotech Discovers a Lung Cancer Drug Candidate
Nature reported that researchers built a 'virtual biotech' of up to 37,000 AI agents coordinated by a simulated chief scientist, which independently navigated the drug discovery pipeline and identified a promising lung cancer drug candidate now entering experimental validation — the most complete AI-led drug discovery arc to reach a major journal.
- Agent hierarchy mirrored actual biotech team roles: target identification, compound screening, ADMET prediction, and regulatory review
- The team claims the discovery timeline compressed from years to weeks — no peer-reviewed external replication has yet been published
- Lung cancer is the leading cause of cancer death globally; the specific target identified has not been publicly named pending IP filing