Ligand Pro, founded by Skoltech professors and a Skoltech Ph.D. student, unveiled Matcha, an AI-powered molecular docking model designed to speed virtual drug screening. The company claims Matcha performs docking about 30 times faster than large co-folding models in the AlphaFold class while maintaining accuracy and “physical correctness” based on its evaluation metrics. The work targets early-stage drug development bottlenecks—where docking throughput often limits how quickly teams can triage candidate libraries. If replicated across diverse target classes and scoring functions, the approach could shorten discovery cycles and improve iteration rates on hit-to-lead programs.