What could robots learn from the videos we already have?
Help answer that question in NVIDIA’s Video to Data (V2D) Challenge. Turn human demonstrations into robot skills across three tracks: 4D Reconstruction, Robotic Grounding, and Egocentric Video-to-Policy.
Join the challenge for a chance to:
- Earn recognition: compete for Track Winner, Runner-Up, and Industry Innovation awards.
- Bring your research to CoRL 2026: top teams are invited to present at the tutorial, with remote participation available.
- Give your solution lasting impact: top solutions receive evaluation at scale and invitations to integrate their code into a maintained repository.
- Help author the technical report: every team with a valid final submission is invited to contribute as a co-author; awardees serve as core authors.
Registration is open now. The challenge kicks off September 21, 2026:
- Learn more: #corl2026 | NVIDIA Robotics | 10 comments
- Get help and ask questions: visit the NVIDIA Omniverse Discord
#v2d-challengechannel for challenge support. - Register for the challenge here: https://nvidia-isaac.github.io/video_to_data/v2d_challenge/
- Attend the CoRL workshop: From Human Videos to Dexterous Robot Capability — CoRL 2026 Tutorial