Benchling announced integration of NVIDIA BioNeMo models and NIM microservices into its AI platform, initially enabling users to run OpenFold2 structure predictions inside Benchling and promising future capabilities such as generative modeling and property prediction. The partnership aims to let scientists apply biomolecular models directly to experimental data without moving between tools. Benchling said integration will scale access to protein structure prediction and, over time, add generative design and property models to accelerate discovery workflows. Nvidia’s NIM microservices allow running pre‑packaged AI models at scale, addressing infrastructure hurdles for lab teams. The deal reflects growing convergence of enterprise lab software and provider‑grade biomolecular AI—centralizing design, prediction and data management to shorten design‑make‑test cycles.