Medra introduced AI Experimentalist, a scientific reasoning layer that translates high-level research goals into end-to-end experimental workflows, as part of a new collaboration with DARPA. The platform is designed to connect literature review, wet-lab execution, data analysis, and protocol refinement, aiming to reduce the bottleneck between hypothesis generation and validated experiments. Medra said AI Experimentalist can propose experimental cascades, including parallelization steps and workflow optimizations, then feed results into the next run. The system is positioned as “physical AI,” using on-site lab hardware and sensors to make decisions during execution rather than stopping at simulation. Partners can access the reasoning layer through Medra’s physical AI labs deployed at customer sites or through remote access to Medra Lab 001. Medra highlighted a 24/7 operational model for running experiments and iterative optimization loops. The DARPA-linked push signals increasing agency interest in automating experimental validation at scale, not only improving in-silico discovery pipelines.