TechBio 3.0 closes the drug discovery loop with multimodal AI and generative chemistry - npj Drug Discovery

Rising drug development costs and flat approval rates have pushed researchers to look for faster ways to identify safer candidates earlier. The piece notes that bringing a new drug to market now exceeds $2 billion and often takes more than 12 years, while nearly 90% of clinical candidates fail to win approval; safety concerns contribute across stages, including late-stage trials where investments are highest. The article explains why predictions based only on molecular structure struggle to capture complex biological behavior, including off-target interactions, tissue accumulation, metabolism into reactive intermediates, and stress responses. It reviews the evolution of “TechBio” and describes TechBio 1.0 (2015–2019) using machine learning for property prediction, highlighting persistent gaps that later generations aim to close with multimodal AI and generative chemistry.