AstraZeneca details AI’s expanding role in biologics drug discovery and R&D acceleration

AstraZeneca says AI is now a core part of its biologics research and development, embedded across design, make, test, and analyze stages to compress cycle times and increase productivity. Puja Sapra, SVP and head of R&D biologics engineering and oncology targeted discovery, describes a build–measure–learn loop in which AI models generate or prioritize candidate molecules, enabling scientists to focus lab resources on the top-ranked options and iterate faster. According to Sapra, this reduces dead ends and allows the company to pursue disease targets previously considered untreatable.

AI is also being applied to the discovery of next-generation biologics that can act on multiple targets simultaneously or deliver therapies precisely to specific cells, which requires optimizing across many variables at once. Sapra says AI-driven models can help identify which targets to prioritize and balance potency, stability, manufacturability, and safety when designing multi-specific molecules.

To consolidate data and close the loop between prediction and experiment, AstraZeneca is building a 'lab of the future' facility in Kendall Square, Cambridge, Massachusetts. The site will combine AI-driven predictions, robotic automation, and continuous data generation so that experimental outcomes feed back into models on a weekly basis. Sapra likens the setup to a self-driving car stack, with AI making predictions, robots executing experiments, and instruments generating data that is immediately piped back for model refinement.

Sapra emphasizes that the company’s datasets—proprietary and multimodal, spanning molecular structures, binding measurements, safety profiles, and manufacturing outcomes—are a key differentiator. She notes that deep screening technologies are being used to generate additional datasets at volume to refine and validate frontier AI models. McKinsey is cited estimating that generative AI, combined with other computational tools, could cut drug discovery timelines by as much as 50%.

Looking further ahead, AstraZeneca is working toward 'de novo' design, where AI generates entirely new protein sequences tailored to desired drug properties, predicts safety and behavior in the body, and informs manufacturability. Sapra says the field is making progress toward AI-designed biologics that proceed all the way to clinical candidates, but she highlights three prerequisites: richer and more standardized training data across the industry, robust evaluation benchmarks for AI-generated candidates, and teams skilled at the intersection of machine learning and biology.

Safety prediction is singled out as especially consequential yet under-discussed, according to Sapra. She describes the company’s use of virtual clinical trials—advanced cell systems and micro-scale organ models paired with AI—to generate enhanced biological signals without traditional testing bottlenecks. These systems are positioned as a critical missing piece in closing the loop between AI-generated designs and clinical-ready candidates.

Sapra also points to a shift toward agentic AI systems that can simultaneously generate molecule candidates and predict efficacy and safety, connecting disease-level insights directly to molecule design and bridging previously separate data silos.