XDOF, just three months out of stealth, is in talks for a Series B at a $1.2B valuation

XDOF, a robotics data startup that emerged from stealth less than three months ago, is in late-stage negotiations to raise a Series B round at a valuation of roughly $1.2 billion, according to multiple people with knowledge of the discussions. The round is expected to be led by 8VC. Terms are not final and could shift, and the total amount being raised or whether the valuation includes new funding has not been disclosed.

The company, co-founded in 2024 by UC Berkeley researchers Philipp Wu, who serves as CEO, and Fred Shentu, the CTO, builds data pipelines, collection tools, and annotation systems for frontier AI labs and robotics firms. XDOF positions itself as an outsourced data-supply chain for the robotics industry, which lacks the vast public datasets that powered large language models. Its technology stems from GELLO, a low-cost teleoperation system the founders developed that lets human operators remotely control robotic arms to generate training data.

XDOF’s rapid growth has driven investor interest. The startup reported annualized revenue approaching $50 million, which prompted venture capitalists to approach the company about a new round even though XDOF had not planned to raise again so soon. In June, the company closed a $70 million Series A with participation from Thrive Capital, Andreessen Horowitz, Lux, and Spark Capital, as previously reported.

The startup is partnering with UC Berkeley’s AI Research lab to release a dataset it calls ABC, which it says will be the largest collection of high-quality robot training data ever assembled. To gather that data, XDOF combines remote robot teleoperation with human collectors who wear sensors to record everyday tasks such as folding clothes and flattening boxes. The company plans to hire and train teams of data collectors globally, including teleoperators who steer robots remotely and egocentric operators who wear body sensors to capture movement data.

Investors have described XDOF as the Scale AI or Mercor for physical robotics, a reference to the data-labeling firms that fueled the AI boom. Unlike large language models trained on internet text, physical robots lack an equivalent real-world dataset, making data collection a key bottleneck for building general-purpose machines. XDOF previously told TechCrunch that it works with 20 customers, including several frontier AI labs.

Other startups pursuing real-world data collection for robot training include Mecka AI, as well as human-data platforms expanding beyond language models, such as Scale AI and Micro1.

More company and startup news from TechManNews.