Amazon launches Bio Discovery
Amazon introduced Bio Discovery, an agentic AI product that brings lab‑in‑the‑loop drug discovery workflows to a broader researcher audience. The service aims to combine agentic AI with laboratory processes to accelerate research and discovery access. (x.com)
Drug discovery starts with a loop: design a molecule on a computer, test it in a lab, then use the result to design the next one. Amazon Web Services said on April 14 it is packaging that cycle into a new product called Amazon Bio Discovery. (aboutamazon.com) Amazon Bio Discovery is an Amazon Web Services application for early-stage research, where scientists look for antibodies or other drug candidates worth testing further. Amazon said the service gives researchers access to biological foundation models, or artificial intelligence systems trained on large biology datasets, inside one workflow. (aws.amazon.com) Amazon said researchers can use the product without writing code, and can move from target selection to molecule design, screening, optimization, and experiment tracking in the same system. Reuters reported the launch on April 14 and said the product is aimed at speeding early-stage drug discovery. (reuters.com) The lab step is the bottleneck Amazon is trying to address. Amazon said the software can hand off designs to contract research organizations for wet-lab testing, then feed the results back into the next round of model-guided design. (aws.amazon.com) That approach reflects a wider shift in drug research over the past two years, as cloud vendors and drugmakers moved from offering standalone models to offering full design-build-test-learn pipelines. Amazon had already been selling HealthOmics tools for bioinformatics workflows and access to models such as AlphaFold and ESMFold before this launch. (aws.amazon.com) Amazon’s new product also pushes the company deeper into a crowded field that includes specialist artificial intelligence drug-discovery companies and large cloud rivals selling life-sciences infrastructure. Axios reported the launch as Amazon’s entry into a more competitive race to cut the cost and time of early drug development. (axios.com) Amazon said the service includes a catalog of more than 40 biological foundation models and lets customers bring in models they built or licensed elsewhere. Its public product materials list tools for antibody structure prediction, humanization, aggregation prediction, and binding analysis. (aws.amazon.com) Amazon is also using outside research groups as reference customers. The company’s product page includes endorsements from scientists at the Broad Institute and Voyager Therapeutics, and trade publications reported a partnership with Memorial Sloan Kettering focused on antibody design for pediatric cancers. (aws.amazon.com; mediapost.com) The pitch is straightforward: put the computer models, the experiment log, and the lab handoff in one place so more biology teams can run the same loop faster. Whether that shortens the path to an actual medicine will still depend on what happens after the first promising hit comes back from the lab. (aboutamazon.com)
