Insilico Achieves 99% Cost Savings in Drug Candidate Discovery Using AWS

Historically, taking a drug from target hypothesis to candidate validation took about 4.5 years and $674 million. Insilico Medicine built ML-powered PandaOmics (target identification via biomarker discovery) and Chemistry42 (compound design/optimization using ML and physics-based methods) on AWS, using Amazon EC2 for scalable GPU compute and Amazon S3 for global data hosting and cross-border team collaboration among 150+ collaborators. Using this stack, Insilico took a fibrosis drug candidate from target discovery to compound validation in under 18 months for $2.6 million — roughly a 99% cost reduction versus the historical $674 million baseline and about 3 years faster than the industry average.

AWS is Amazon's cloud platform, used across pharma for genomics compute, SageMaker/Bedrock model training, and healthcare-and-life-sciences-specific services like Amazon HealthOmics.

This sits in the Build stage of the AI use-case lifecycle: Named examples from the foundation-model labs and platforms pharma teams use to actually construct, train and generate the model or application.