Two AI‑driven advances signaled fast adoption of machine learning across discovery and translational pipelines. An MIT team released BoltzGen, an open‑licensed generative model that designs therapeutic modalities (nanobodies, mini‑binders, peptides) across proteins and nucleic acids and is backed by a large validation network. Early wet‑lab collaborations reported nanomolar affinities for select designs. Separately, Guardant Health and Zephyr AI announced a strategic partnership to combine Guardant’s multimodal molecular datasets with Zephyr’s analytics to accelerate genomic biomarker discovery for oncology drug development and response prediction. Guardant positioned the collaboration within its Infinity AI capabilities to enhance precision‑oncology biomarker generation. Together, these efforts illustrate parallel industry moves: foundation models for de novo therapeutic design and applied AI to extract clinical signal from large molecular datasets — both aiming to compress R&D timelines and broaden the druggable landscape.