The 2026 edition in Austin, Texas, presents original research on learning representations for robotic perception and control; robot foundation models and general-purpose knowledge systems; imitation learning, including behavioural cloning and inverse reinforcement learning; reinforcement learning for physical robots; model-based and model-free control and decision-making; combining learning and planning, including task and motion planning; probabilistic learning and uncertainty; automatic robotic data generation; multimodal perception, sensor fusion and robot vision; human-robot interaction and instruction by language and gestures; learning for hardware design; robot safety and alignment; applications in manipulation, navigation, locomotion, driving and flight; robot systems and sensors for data-driven approaches; theoretical foundations including generalisation and uncertainty quantification; video and latent world models; and benchmarks and datasets. The main conference runs from 9 to 11 November 2026, with keynotes, early-career keynotes, a special panel, papers, demos and late-breaking results, followed by 32 accepted workshops on 12 November 2026. Sessions take place at the Hilton Austin (presentations and keynotes on days 1 and 2), the JW Marriott Austin (posters, exhibits, lunches and breaks) and Bass Hall at The University of Texas at Austin (reception and day-3 keynote).