Follow the latest coverage, related explainers and connected technology stories.
A facility in Narashino, Japan, has begun collecting training data for physical AI using 35 humanoid robots as part of a consortium comprising five companies. The facility relies on remote control, motion recording, and labeling successful and failed states, with the goal of building models that can be used in factories and various industries.
Less than three months after emerging from stealth, startup XDOF is negotiating a Series B funding round led by 8VC at a valuation of approximately $1.2 billion. The company collects the remote-operation and motion data needed to train versatile robots.
Encord, in collaboration with Zander Labs, is testing the use of brain-wave data alongside video and sensor data to create richer training data for robots. The experiment reflects a broader problem: the scarcity of real-world data and the high cost of producing it compared with training traditional artificial intelligence models.