Deepen Refinery adds ROVR’s 90,000-plus hours of driving data

Deepen AI has added ROVR’s real-world multimodal driving dataset to Deepen Refinery, expanding the marketplace’s supply of verified sensor data for autonomous vehicles, robotics and Physical AI. The move is meant to help developers find, evaluate and use large-scale data faster, while giving ROVR a broader channel to commercialize its collection.

Why it matters: - Physical AI teams need more than raw volume. They need verified data with clear provenance, usable formats and enough context to trust it in training and deployment. - The addition gives autonomous vehicle and robotics developers access to a larger real-world dataset across multiple regions and sensor types. - ROVR gains a distribution channel into a broader ecosystem of AI teams without building its own marketplace infrastructure.

What happened: - Deepen AI announced that ROVR’s real-world multimodal driving data is now available through Deepen Refinery. - Deepen Refinery is Deepen AI’s neutral marketplace for verified data used in autonomous vehicles, robotics and other Physical AI applications. - The dataset includes more than 90,000 hours of driving data. - The announcement was made Oct. 5, 2026, in Santa Clara, California.

The details: - ROVR’s automotive dataset includes forward-facing LiDAR and camera data. - Each ROS2 bag contains LiDAR, imagery, IMU and GNSS data, along with calibration parameters. - The dataset covers driving environments across the Americas, Europe and Asia. - The underlying data pool is continuously refreshed as new data is collected. - Deepen Refinery is designed to catalog, structure, enrich, validate and make datasets easier to discover and evaluate. - The platform preserves data provenance and dataset characteristics. - Deepen Refinery can deliver data in formats aligned with modern AI and robotics workflows, including ROS2, MCAP, HDF5, nuScenes and LeRobot. - The marketplace also provides a foundation for data validation and quality assessment before datasets enter development pipelines. - The platform connects data contributors with organizations building autonomous vehicles, robotics and other intelligent systems.

Between the lines: - The deal reflects a wider shift in Physical AI from collecting more data to making data usable, comparable and trusted. - Deepen AI is positioning Refinery as infrastructure that sits between fragmented data sources and development teams. - The structure favors two groups at once: contributors that want to monetize data and developers that want faster access to verified datasets. - Deepen AI CEO Mohammad Musa said Physical AI teams need data they can understand, trust and use, and that Deepen Refinery adds the infrastructure around ROVR’s dataset. - ROVR co-founder Gary Ling said the partnership helps make its data accessible to teams building autonomous vehicles, robotics and Physical AI.

What’s next: - Deepen AI is likely to keep expanding Refinery’s catalog as more specialized contributors join the marketplace. - ROVR’s continuously refreshed data pool suggests the dataset will keep growing as new collection comes in. - Demand for diverse real-world data is expected to increase as autonomous systems and robotics become more capable. - Deepen AI says Refinery is designed to serve as a neutral data layer for the Physical AI ecosystem.

The bottom line: - Deepen Refinery is betting that the next bottleneck in Physical AI is not just data collection, but data access, validation and workflow readiness. - More information: Deepen Refinery - More information: ROVR Network - Company page: Deepen AI on LinkedIn