AWS Launches Drug‑Discovery AI

Drug discovery starts with a simple problem: scientists have to search through huge numbers of molecules to find a few that might become medicines. Amazon Web Services said on April 14 it is now selling a tool called Amazon Bio Discovery to speed that early search. (aboutamazon.com) Amazon Bio Discovery is an artificial intelligence application that lets researchers run early-stage drug research workflows without writing code, according to Reuters and Amazon Web Services. The product launched Tuesday, April 14, at the Amazon Web Services Life Sciences Symposium in New York. (usnews.com, aws.amazon.com) Amazon Web Services said the software gives scientists access to specialized biological foundation models, which are large artificial intelligence models trained on biological data such as protein structures and sequences. The company said researchers can use those models to generate and test ideas for new drug candidates faster. (aboutamazon.com, aws.amazon.com) The launch extends Amazon Web Services’ existing HealthOmics business, which already sells managed computing and storage for genomics and bioinformatics work. On its HealthOmics site, Amazon Web Services says the service supports models including AlphaFold, ESMFold, and NVIDIA BioNeMo for production-scale pipelines that can process millions of candidates. (aws.amazon.com, aws.amazon.com) Drugmakers and technology companies have spent the past two years pushing artificial intelligence deeper into research labs, especially in the first stage of drug development, when teams decide which biological target and which molecule to pursue. Amazon Web Services is pitching its product as a combined software-and-computing service rather than a single model. (usnews.com, aws.amazon.com) Amazon Web Services said the system can orchestrate model runs, simulations, and downstream analysis inside one managed environment. That matters in practice because drug research teams often stitch together separate tools for chemistry, biology, and high-performance computing. (aws.amazon.com, aws.amazon.com) The company has been laying groundwork for this launch inside its life sciences unit. In March, Amazon Web Services published a blog post describing a “three-week drug discovery agent” built with its Kiro coding tool and Amazon Bedrock AgentCore, and another post on using Kiro with HealthOmics workflows. (aws.amazon.com, aws.amazon.com) Amazon Web Services is also leaning on customer case studies to make the sales pitch. In July 2025, the company said Sonrai cut research timelines by 70 percent and reduced some experimental costs by as much as 98.6 percent after moving workflows onto HealthOmics. (aws.amazon.com) The bet is that more of drug discovery will be bought as cloud infrastructure plus ready-made artificial intelligence tools, not built from scratch inside each lab. Amazon Web Services is trying to make that model part of its healthcare business before more of the early research stack gets locked up by rivals. (aboutamazon.com, aws.amazon.com )