Google announced an AI model that proposed and validated a novel cancer drug combination, marking a high‑visibility step for large‑scale foundation models applied to small‑molecule and combination discovery. The effort used a 27‑billion‑parameter model family and moved from in silico proposal to experimental validation, signaling growing confidence in generative AI for lead identification. In parallel, Foundation Medicine expanded its FoundationInsights analytics with an AI partnership to deliver natural‑language search and programmatic analysis to biopharma customers. The deal aims to make large, de‑identified genomic datasets more actionable for translational research and precision oncology decisioning. Taken together, the announcements show pharmaceutical R&D groups and platform companies doubling down on AI to shorten discovery timelines and improve target selection, while also prompting questions about validation, reproducibility and regulatory expectations for AI‑generated candidates.