Researchers in China introduced DrugCLIP, an AI framework that screens millions of small molecules against thousands of protein targets in hours — a reported acceleration up to ten million times versus traditional virtual screening. The tool combines large‑scale protein–ligand modeling with high‑throughput inference to prioritize candidates rapidly, potentially compressing early discovery timelines. Authors say DrugCLIP could reshape hit identification, but integration with experimental follow‑up and ADMET profiling remains necessary to convert computational hits into validated leads.