Less than three months after emerging from stealth, XDOF, a startup that collects real-world remote control data to train general-purpose robots, is in late-stage negotiations to raise a Series B led by 8VC at a valuation of about $1.2 billion, people familiar with the deal said.
XDOF was co-founded in 2024 by University of California, Berkeley researchers Philip Wu (CEO) and Fred Shentu (CTO). In June, TechCrunch reported on the startup’s $70 million Series A with participation from Thrive Capital, Andreessen Horowitz, Lux and Spark Capital. XDOF did not plan to raise again immediately after that round. But the company’s rapid growth (annual sales are approaching $50 million) has led venture capitalists to approach it for new rounds, the people said.
TechCrunch was unable to determine whether the total capital raised or the valuation includes new funding. Terms and conditions are not final and are subject to change.
XDOF and 8VC did not respond to requests for comment.
The startup aims to build data pipelines, collection tools, and annotation systems that cutting-edge AI labs and robotics companies can’t easily build in-house, essentially acting as an outsourced data supply chain for the robotics industry.
As a doctoral student, Wu researched how robots learn from large datasets. One of the big hurdles for his research was the lack of “large-scale data to work with,” he told TechCrunch in June.
There he collaborated with Shentu on a project called GELLO. GELLO is a low-cost teleoperation system that allows a human operator to remotely control a robotic arm and generate training data. Their research led to seminal papers in robotics.
This research formed the basis of XDOF. XDOF is now being referred to by investors as the Mercor of Scale AI or physical robotics, referring to the data labeling giant that helped fuel the AI boom. Unlike LLMs, which were initially trained across the Internet, physical robots do not have comparable real-world datasets to draw from, making data collection a critical bottleneck in building general-purpose machines.
XDOF is partnering with the University of California, Berkeley’s AI Research Institute to release what is believed to be the largest collection of high-quality robot training data ever collected, called ABC.
To capture this data, XDOF combines the remote control of remote robots with human collectors who wear sensors and record daily tasks such as folding clothes or flattening boxes.
The company plans to hire and train a team of data collectors around the world, including remote operators who control the robots remotely and self-centered operators who wear body sensors to collect movement data.
XDOF previously told TechCrunch that it is already working with 20 customers, including several frontier AI labs.
Other startups trying to collect real-world data for robot training include Mecka AI and human data platforms like Scale AI and Micro1 that are expanding beyond LLM.
If you make a purchase through links in our articles, we may earn a small commission. This does not affect editorial independence.
