XDOF, a startup that collects real-world teleoperation data for training general-purpose robots, is in late-stage talks to raise a Series B at a valuation of roughly $1.2 billion, led by 8VC — less than three months after emerging from stealth, TechCrunch reported, citing people familiar with the matter.
The company was co-founded in 2024 by UC Berkeley researchers Philipp Wu (CEO) and Fred Shentu (CTO). TechCrunch reported on a $70 million Series A in June with participation from Thrive Capital, Andreessen Horowitz, Lux and Spark Capital. XDOF had not planned to raise again so soon, but its rapid growth — annualized revenue is approaching $50 million — prompted investors to approach it about a new round. The total being raised and whether the valuation includes the new money were not clear, and terms could still change.
XDOF builds the data pipelines, collection tools and annotation systems that frontier AI labs and robotics companies cannot easily build themselves — an outsourced data supply chain for physical AI. Investors describe it as the "Scale AI for physical robotics": unlike large language models, which trained on much of the internet, robots have no equivalent real-world dataset, making data collection the critical bottleneck for general-purpose machines.
The company's research roots include GELLO, a low-cost teleoperation system that lets human operators steer robotic arms remotely to generate training data, and an influential paper that grew out of that work. XDOF is partnering with UC Berkeley's AI Research lab to release what it believes is the largest collection of high-quality robot training data ever assembled, dubbed ABC. Data capture combines remote teleoperation with human collectors who wear sensors to record everyday tasks such as folding clothes and flattening boxes, and the company plans to hire and train collector teams worldwide. It says it already works with about 20 customers, including several frontier AI labs.
