XDOF, just three months out of stealth, is in talks for a Series B at a $1.2B valuation
TechCrunch Marina Temkin
XDOF is in talks for a new round at about $1.2 billion, just months after coming out of stealth. Investors are betting real robot data is becoming the new bottleneck.
Based on reporting by TechCrunch, Marina Temkin — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
XDOF, the robot-data startup that only came out of stealth less than three months ago, is now in late-stage talks to raise a Series B at a valuation of about $1.2 billion, according to people familiar with the deal. 8VC is said to be leading the round. The terms aren’t final yet, and the total amount hasn’t been learned.
The speed is the part that jumps out. TechCrunch reported XDOF’s $70 million Series A in June, and the company wasn’t even planning to go back to the market this soon. But revenue growth changed the mood fast: people familiar with the company say its annualized revenue is approaching $50 million, which is enough to make VCs start chasing it instead of the other way around.
XDOF was founded in 2024 by UC Berkeley researchers Philipp Wu, who is CEO, and Fred Shentu, who is CTO. The company is trying to build the unglamorous plumbing behind robot training: data pipelines, collection tools and annotation systems that frontier AI labs and robotics companies would rather not build themselves. That makes it something like an outsourced supply chain for physical AI.
The company’s roots go back to GELLO, a low-cost teleoperation system Wu and Shentu built so a human operator could remotely control a robotic arm and generate training data. Their work led to an influential robotics paper, and now XDOF is trying to scale that idea. It is also partnering with UC Berkeley’s AI Research lab on ABC, which it says will be the largest collection of high-quality robot training data ever assembled.
That data comes from a mix of remote teleoperators and human collectors wearing sensors while they do everyday tasks like folding clothes and flattening boxes. XDOF says it already works with 20 customers, including several frontier AI labs. And the broader market is starting to look crowded: Mecka AI is chasing the same real-world data opportunity, while Scale AI and Micro1 are stretching beyond LLMs into human data.
My take — AI-written commentary, not fact-checked reporting
This is the kind of startup deal that says more about the market than the company. Everyone spent years pretending model scale was the only scarce thing; now the money is flowing toward the messier part, where robots actually need data from the real world. That’s a healthier bet than another round of chatbot confetti.
Read more about this at: TechCrunch