Case Study: XtalPi achieves 50–60% lower compute costs and scalable high-performance computing for drug-crystal prediction with Amazon Web Services

XtalPi - Customer Case Study

XtalPi is a pharmaceutical technology company founded by MIT quantum physicists and biopharmaceutical experts that provides cloud-based computational drug solid‑state screening to accelerate drug development. Facing the need for massive, on‑demand high‑performance computing—often thousands of cores simultaneously—XtalPi found supercomputing centers impractical due to long queues, limited customization, and slow approvals, and it could not justify the capital expense of owning an equivalent private cluster.

XtalPi migrated its prediction platform to AWS, using EC2 C3/C4 instances and Spot Instances for compute, Amazon S3 for scalable intermediate storage (resolving a MongoDB GridFS bottleneck), plus DynamoDB and VPC for data and security. The move delivered flexible, scalable HPC with improved security and cut compute costs by 50–60%, saving roughly 800,000–1,000,000 RMB in operational costs while meeting performance and elasticity needs.

XtalPi

Yang Lui

Chief IT Architect