A robotics startup raises $30M for the humanoid training-data shortage
From the Robots digest for 2026-10-22
What happened: Forbes (John Koetsier, ~4:46 PM ET October 22) reports that a robotics startup has raised $30 million to attack the robot training-data shortage, with the piece framed around the question the whole lane is built on — how to make humanoid robots smart by giving them more training data. The captured summary does not name the company or break down how the money is split between human task-data collection and synthetic or simulated data.
What's genuinely new: Development of the training-data lane — the record's facilities and methods entries (DexForce's simulation training, Beijing's Shijingshan data-training center, Japan's J-HRTI Kanto Data Factory, the gig-style human task-data collection workforce) now gain a venture-capital bet aimed directly at the data shortage itself. The $30 million is the lane's first fresh raise figure in this window; the company behind it and the shape of its data operation are not recoverable from the capture, and the raise is a single-source company-claimed figure pending the round's own paperwork.
Watch: The startup's identity and approach (human-collected versus synthetic data); whose humanoids consume its data; whether the raise shows up in facility openings or a named training product.
Links: Forbes