Robot-data startup XDOF reportedly nears a $1.2 billion valuation

Robot training-data company XDOF is in late-stage talks for a Series B led by 8VC at a valuation of about $1.2 billion, according to people cited by TechCrunch. The terms remain unsettled, and the report does not identify the amount being raised or whether the valuation includes the new capital. XDOF and 8VC did not comment.

The talks come less than three months after XDOF disclosed a $70 million Series A. TechCrunch reports that annualized revenue is approaching $50 million and that the company works with 20 customers, including several frontier AI laboratories. Those figures help explain why investors approached it sooner than expected.

XDOF builds tools and collection pipelines for the physical data used to train general-purpose robots. Remote operators steer robotic arms to record tasks, while people wearing sensors capture movements such as folding clothes and flattening boxes. The company grew from GELLO, a low-cost teleoperation research project created by co-founders Philipp Wu and Fred Shentu at UC Berkeley.

Unlike language models, robots cannot draw a broad representation of physical interaction from web text alone. XDOF is also partnering with Berkeley’s AI research lab on a collection called ABC, which it says will be the largest high-quality robot-training dataset. The financing report remains provisional until a round closes.