MIT and Recursion’s Boltz‑2 model demonstrated rapid, accurate binding‑affinity predictions, reporting affinity calculations in seconds and matching top performance in recent CASP benchmarking. The release of high‑performance, permissively licensed models is accelerating adoption of physics‑aware AI in small‑molecule discovery and may reduce reliance on costly free‑energy simulations. Concurrently, the Chan Zuckerberg Initiative expanded collaboration with NVIDIA to scale its Virtual Cell platform, bringing GPU acceleration and infrastructure to large biological models. The partnership aims to harmonize massive single‑cell datasets and make AI‑driven virtual cells accessible to the community. These twin developments—faster affinity prediction and scalable biological modeling—signal a practical step toward AI‑assisted lead identification and mechanistic simulation in drug discovery workflows.