Nexdata to show Physical AI data tools at IROS 2026

Nexdata will exhibit its Physical AI data resources and services at IROS 2026 in Pittsburgh from Sept. 27 to Oct. 1. The company will highlight teleoperation, UMI, ego-centric and simulation data for robot learning and model development.

Why it matters: - Physical AI systems need large, varied datasets to improve perception, decision-making and interaction in real-world and simulated environments. - Nexdata is positioning its data factory and services for robot learning teams that need off-the-shelf data, custom collection and validation support.

What happened: - Nexdata will attend IROS 2026, the IEEE/RSJ International Conference on Intelligent Robots and Systems, in Pittsburgh, Pennsylvania, from Sept. 27 to Oct. 1. - The company will exhibit at Booth #640. - Nexdata will present Physical AI data resources and services focused on teleoperation, UMI, ego-centric and simulation data. - Minji Bai will give a presentation titled “Nexdata Physical AI Data Solution” on Tuesday, Sept. 29, from 3:45 p.m. to 4:00 p.m. - Visitors can meet the Nexdata team at the booth throughout the conference.

The details: - Teleoperation offerings include 10,000+ hours of dexterous hand teleoperation datasets and 266,000+ images for robot vacuum cleaners and robotic lawn mowers. - UMI services cover bespoke data collection across different scenarios and tasks, with monthly capacity of 50,000+ hours. - Ego-centric datasets include 135,000+ hours across single-camera, dual-camera and multi-camera setups. - Simulation assets include 288 million 3D models and scenes, plus scan-to-sim services for Physical AI simulation development. - Nexdata operates an 8,000 m² Physical AI data factory. - The facility includes collection environments for home, retail, warehouse, industrial, office and other settings. - The platform supports humanoid robots, robotic arms and dexterous hands. - Collection tasks include grasping, placing, carrying, sorting, organizing and assembly. - Nexdata also offers off-the-shelf datasets, bespoke data collection, data annotation and model validation. - The company says the presentation will explain its Physical AI data approach and its support for robot learning and model development. - Nexdata says its services also support Generative AI, Conversational AI and other AI applications. - More information is available on Nexdata's website. - Nexdata's social channels include LinkedIn, Facebook, YouTube and X.

Between the lines: - Nexdata is using IROS to reach robotics researchers and builders who need large-scale training data, not just software tools. - The mix of teleoperation, ego-centric and simulation assets suggests the company wants to cover multiple stages of Physical AI development, from data capture to model testing. - The booth and talk create a sales-and-demo moment around a category that still depends heavily on data availability and quality.

What's next: - Nexdata will demo its datasets and Bespoke services at Booth #640 during the conference. - The Sept. 29 presentation will give attendees a closer look at the company’s Physical AI data workflow. - Interest from robotics teams at IROS could shape follow-up projects and service demand after the event.

The bottom line: - Nexdata is betting that Physical AI progress will be driven by better data pipelines, and it is using IROS 2026 to make that case to robotics developers.