A year after scientists first sketched antibodies from scratch with AI, the technology is edging toward the clinic. Academic groups and startups report de novo antibodies—both nanobodies and full-length molecules—with drug-like potency, specificity and manufacturability, including against hard targets such as GPCRs. Advances in modeling flexible antibody loops, including updates inspired by AlphaFold and new generative tools like BoltzGen, are compressing design cycles and slashing wet‑lab screening. Investors and drugmakers see a path to faster, cheaper biologics in a market where antibody therapies already generate tens of billions in sales, though real-world efficacy, safety, and regulatory validation remain the next hurdles. Expect the first AI-designed antibody candidates to enter human trials as early as the next development window, setting the stage for a new phase in biologic drug R&D.
Related article:
— Improved de novo antibody design (Bennett et al., Nature, 2025)
