AI reshapes biologic drug discovery

AI is becoming core infrastructure for biologic drug discovery, where scientists must search vast molecular spaces for candidates that bind to the right target, remain stable in the body and can be manufactured at scale. AstraZeneca’s Puja Sapra says the company now uses computational tools across design, making, testing and analysis, with AI helping generate or prioritize molecules before lab resources are committed.

The approach relies on a build-measure-learn loop that narrows testing to top-ranked candidates and feeds experimental results back into models. McKinsey estimates that generative AI, combined with other computational tools, could cut drug discovery timelines by as much as 50%. AstraZeneca says its proprietary multimodal datasets, including molecular structures, binding measurements, safety profiles and manufacturing outcomes, help fine-tune models for biologics research.

AstraZeneca is also developing a “lab of the future” in Kendall Square, Cambridge, Massachusetts, where AI, robotic automation and instruments would operate in a closed-loop discovery system. Sapra says automated systems could eventually make and evaluate thousands of molecular interactions on a weekly basis, generating AI-ready data beyond traditional workflows.

The longer-term goal is de novo design, where AI generates new protein sequences with desired drug properties from scratch. Sapra identifies richer standardized data, stronger benchmarks, interdisciplinary talent and better safety prediction as key requirements, with scientists providing oversight as agentic AI systems connect disease insights to molecule design.