Tempus AI partnered with Daiichi Sankyo to develop proof-of-concept AI models aimed at improving success rates for novel oncology antibody-drug conjugates (ADCs). Under the agreement, Daiichi Sankyo will access Tempus’ foundation models and AI expertise, including PRISM2. The collaboration focuses on biomarker discovery and patient stratification, with an emphasis on benchmarking control arms for future clinical trials. Tempus said the models will support translational workflows by linking pathology-derived information with broader datasets. Tempus has expanded its collaboration footprint recently, including work with Merck on precision medicine biomarker discovery and partnerships to build real-world registries. For Daiichi Sankyo, the ADC emphasis aligns with ongoing portfolio efforts to improve targeting and reduce trial risk. The deal reflects a broader push to apply foundation models to patient selection and trial design—not just retrospective analysis—while tightening the evidentiary loop from biomarkers to clinical outcomes.