XDOF, a robotics data startup that emerged from stealth less than three months ago, is reportedly in advanced talks to raise a Series B at a valuation of around $1.2 billion. The round is said to be led by 8VC.
Founded in 2024 by UC Berkeley researchers Philipp Wu and Fred Shentu, the company builds the data pipelines, collection tools, and annotation systems needed to train general-purpose robots. Its model focuses on solving one of robotics' biggest challenges: the lack of large-scale real-world training data.
XDOF first drew attention with a $70 million Series A in June, backed by Thrive Capital, Andreessen Horowitz, Lux, and Spark Capital. Since then, the company's growth has accelerated, with annualized revenue approaching $50 million, according to people familiar with the deal.
The startup's work is rooted in teleoperation research, including the GELLO system, which lets humans remotely control robotic arms to generate training data. XDOF now combines remote operation with sensor-based data collection from everyday tasks such as folding clothes and flattening boxes.
In partnership with UC Berkeley's AI Research lab, the company is also preparing ABC, a dataset it says could become one of the largest high-quality robot training collections ever assembled. XDOF has said it already works with 20 customers, including frontier AI labs.
As robotics moves closer to broader real-world use, data infrastructure companies like XDOF may help define the next wave of intelligent machines and shape how robots learn in the years ahead.
