Series
Humanoid Data War 2026
The "Humanoid Data War 2026" series has 7 parts — read them in order from part 1.
Research & TrendsThe Data War: Who Owns Humanoid Robot Data in 2026?
Map the real humanoid robot data ownership landscape: AgiBot World 1M+ trajectories vs Open X-Embodiment vs Physical Intelligence's π0 vs LeRobot — who's winning and why it matters strategically.
Teleoperation: Real-World Robot Data Collection
Compare AgiBot, Unitree XR, ALOHA 2, Mobile ALOHA, and GELLO to choose a real-world teleoperation data stack by budget.
Human Video Mining: Pre-training Robots from Human Video
How Ego4D, EPIC-Kitchens, web video, LAPA, GR00T N1, EgoMimic, and EgoHumanoid turn human video into robot pre-training data.
Synthetic Data Pipelines: From Sim to Real
How to turn a few human demos into thousands of robot trajectories with MimicGen, GR00T-Mimic, RoboCasa, and sim-real co-training.
VLA Data Scaling Laws for Robot Learning
Robot data does not scale by raw demo count alone. This guide explains diversity power laws and when more data becomes wasted spend.
Data Strategy: What Should Small Teams Collect?
A practical playbook for small robotics teams: SO-101 data, LeRobotDataset v2, pretrained VLA fine-tuning, and HF datasets.
Open vs Closed: Licenses, Data Moats, and What's Next
Compare robot dataset licenses, commercial training risk, closed fleet data moats, and the likely open vs closed landscape in 2027.