XDOF: The Robot Data Startup Racing to a $1.2 Billion Valuation Just Months Out of Stealth

In a striking sign of the frenzy surrounding physical AI, 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 about $1.2 billion. This comes less than three months after the company emerged from stealth with a $70 million Series A.

The Whirlwind Ascent

XDOF was co-founded by UC Berkeley researchers Philipp Wu (CEO) and Fred Shentu (CTO) in 2024. The startup’s rapid growth, with annualized revenue approaching $50 million, has prompted venture capitalists to approach it about a new round much sooner than planned. The Series B is reportedly being led by 8VC, with terms still not final and subject to change.

The startup aims to build the data pipelines, collection tools, and annotation systems that frontier AI labs and robotics companies cannot easily build themselves. In essence, XDOF is positioning itself as an outsourced data supply chain for the robotics industry, filling a critical gap in the physical AI ecosystem.

The Data Bottleneck

As a PhD student, Wu was studying how robots learn from large datasets when he encountered a major impediment: the lack of “large-scale data to work with.” Unlike large language models that initially trained on the vast expanse of the internet, physical robots do not have an equivalent real-world dataset to draw from. This makes data collection a critical bottleneck to building general-purpose machines.

Wu teamed up with Shentu on a project called GELLO, a low-cost teleoperation system that allows a human operator to control a robotic arm remotely to generate training data. Their work led to an influential paper in robotics and formed the foundation for XDOF.

The Scale AI for Robotics

Investors now describe XDOF as the Scale AI or Mercor for physical robotics, referencing the data-labeling giants that helped fuel the AI boom. The startup 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.

To capture this data, XDOF combines remote robot teleoperation with human collectors who wear sensors to record everyday tasks like 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.

The Competitive Landscape

XDOF previously told TechCrunch that it is already working with 20 customers, including several frontier AI labs. Other startups attempting to collect real-world data for robot training include Mecka AI, as well as human-data platforms expanding beyond LLMs, such as Scale AI and Micro1.

What This Means for the Industry

The rapid ascent of XDOF reflects a broader recognition that the physical AI revolution requires a fundamentally different infrastructure than its digital counterpart. While language models could ingest the accumulated text of human civilization, robots need to learn from physical interaction with the real world, which must be systematically captured and curated.

The valuation of $1.2 billion for a company just months out of stealth underscores the urgency that investors attach to solving this data bottleneck. As the race to build general-purpose robots intensifies, the companies that control the data pipelines may wield extraordinary power over the future of physical AI.

XDOF’s trajectory also highlights the growing importance of the data supply chain in the AI economy. Just as Scale AI became indispensable to the large language model boom, XDOF and its competitors are positioning themselves as essential infrastructure providers for the next wave of AI innovation.

The coming months will reveal whether XDOF can maintain its blistering pace of growth and whether the market for robot training data will prove as vast as investors currently believe. For now, the startup’s story serves as a powerful reminder that in the AI gold rush, the companies selling shovels and picks often end up being the most valuable of all.

TechTrib.com is a leading technology news platform providing comprehensive coverage and analysis of tech news, cybersecurity, artificial intelligence, and emerging technology. Visit techtrib.com.

Contact Information: Email: news@techtrib.com or for adverts placement adverts@techtrib.com