Researchers published an open-source AI-informed platform, AdaptiveFlow, designed to enable virtual screening at ultra-large scale with reduced computational cost. The approach is described as capable of virtually screening billions of drug-like molecules with a 1,000-fold reduction in computational costs versus existing methods. The work was developed and validated by teams at St. Jude Children’s Research Hospital, University of Pavia, Dana Farber Cancer Institute and Harvard Medical School, and was published in Nature Biotechnology. AdaptiveFlow is positioned to routinely run ultra-large virtual screens by integrating thousands of docking protocols, cloud scaling, and AI/ML methods that prioritize promising chemical subspaces. By lowering the barrier for large-scale in silico campaigns, the platform targets a recurring bottleneck in early-stage discovery where sampling constraints limit hit and lead finding.