XDOF, a startup that collects real-world teleoperation data for training general-purpose robots, is in late-stage talks to raise a Series B funding round at a valuation of approximately $1.2 billion. The round is being led by 8VC, marking a remarkably rapid ascent for a company that emerged from stealth less than three months ago.
The startup was co-founded in 2024 by Philipp Wu, who serves as CEO, and Fred Shentu, who holds the position of CTO. Just three months earlier, in June 2026, XDOF raised a $70 million Series A round with participation from notable investors including Thrive Capital, Andreessen Horowitz, Lux, and Spark Capital.
According to sources familiar with the matter, XDOF had not originally planned to raise additional funding so soon after its Series A. However, the company’s explosive growth trajectory changed those plans. The startup’s annualized revenue is now approaching $50 million, prompting venture capitalists to approach the company about a new funding round.
The foundation of XDOF traces back to research conducted by Wu and Shentu at UC Berkeley. During his time as a PhD student, Wu focused on how robots learn from large datasets but encountered a significant obstacle: the lack of large-scale data available for research purposes. This challenge led him to collaborate with Shentu on GELLO, a low-cost teleoperation system that enables human operators to remotely control robotic arms to generate training data.
XDOF’s business model positions it as a critical infrastructure provider for the robotics industry. The company builds data pipelines, collection tools, and annotation systems that frontier AI labs and robotics companies would struggle to develop independently. In essence, XDOF functions as an outsourced data-supply chain for organizations working on robotics applications.
The startup’s data collection methodology combines two primary approaches: remote robot teleoperation and human collectors equipped with sensors. These human operators wear sensing equipment to record everyday tasks such as folding clothes and flattening boxes, creating a diverse dataset of real-world movements and actions.
XDOF is collaborating with UC Berkeley’s AI Research lab to release the ABC dataset, which the company believes will be the largest collection of high-quality robot training data ever assembled. This partnership underscores the academic roots of the company and its continued connection to cutting-edge research.
The startup has already secured significant commercial traction, working with 20 customers that include several frontier AI labs. This early customer adoption demonstrates the urgent need in the industry for the type of data services XDOF provides.
XDOF faces competition from several players in the robotics data collection space, including Mecka AI, Scale AI, and Micro1. However, the company’s rapid revenue growth and premium valuation suggest it has established a strong position in this emerging market segment.
