via TechCrunch AI
XDOF Nears $1.2B Series B Just Months After Stealth Launch
Less than three months after emerging from stealth, XDOF—a startup specializing in real-world teleoperation data for training general-purpose robots—is in advanced discussions to raise a Series B round at a valuation of approximately $1.2 billion, led by venture firm 8VC, according to multiple sources familiar with the matter.
Founded in 2024 by UC Berkeley researchers Philipp Wu (CEO) and Fred Shentu (CTO), XDOF has quickly become a key player in the robotics data supply chain. In June 2025, TechCrunch reported on the company’s $70 million Series A round, which included participation from Thrive Capital, Andreessen Horowitz, Lux, and Spark Capital. At the time, XDOF had not planned to raise again so soon, but its explosive growth—with annualized revenue now approaching $50 million—prompted investors to approach the company about a new round, sources said.
The final terms of the deal are not yet set, and TechCrunch was unable to confirm the total amount being raised or whether the $1.2 billion valuation is pre- or post-money. XDOF and 8VC did not respond to requests for comment.
XDOF’s mission is to build the data pipelines, collection tools, and annotation systems that frontier AI labs and robotics companies cannot easily develop in-house. In essence, the startup acts as an outsourced data supply chain for the robotics industry, addressing a critical bottleneck: unlike large language models that initially trained on vast swaths of the internet, physical robots lack a comparable real-world dataset.
During his PhD, Wu studied how robots learn from large datasets but was hindered by the scarcity of such data. He teamed up with Shentu to create GELLO, a low-cost teleoperation system that enables human operators to remotely control robotic arms and generate training data. Their research led to an influential paper and laid the groundwork for XDOF.
Investors now describe XDOF as the “Scale AI for physical robotics,” referencing the data-labeling giants that fueled the AI boom. As of late 2025, XDOF is partnering with UC Berkeley’s AI Research lab to release what it claims is the largest collection of high-quality robot training data ever assembled, dubbed the ABC dataset. To compile this data, XDOF combines remote robot teleoperation with human collectors who wear sensors to capture everyday tasks such as folding clothes and flattening boxes.
The startup plans to hire and train teams of data collectors worldwide—including teleoperators who steer robots remotely and egocentric operators who wear body sensors to capture movement data. XDOF has previously confirmed it is working with 20 customers, including several frontier AI labs.
Other startups in this space include Mecka AI, as well as broader human-data platforms like Scale AI and Mercor, which are expanding beyond LLM training into physical robotics data.