Will it run?
Labs

XDOF in talks for Series B at $1.2B valuation three months after stealth exit

By Desmond Okafor Clawpit staff
XDOF in talks for Series B at $1.2B valuation three months after stealth exit

The robotics data infrastructure startup XDOF is in advanced discussions for a second funding round at roughly $1.2 billion, less than three months after its public debut. 8VC is leading the round, according to people familiar with the deal. The company had not planned to raise again so quickly after a $70 million Series A in June, but an annual revenue run rate nearing $50 million pulled investors in.

Founders and technical roots

Philipp Wu, chief executive, and Fred Shentu, chief technology officer, founded XDOF in 2024 while both were researchers at UC Berkeley. As a doctoral student, Wu studied how robots learn from large datasets and hit a wall: there simply wasn't enough data at the required scale. The solution he built with Shentu was GELLO, a low-cost teleoperation system that lets a human operator control a robot arm remotely and generate training data. The work produced an influential robotics paper and became the technical foundation for XDOF.

Business model: "Scale AI for physical robotics"

Investors describe XDOF as the physical-world counterpart to Scale AI or Mercor, the data-labeling platforms that powered the LLM boom. The fundamental difference: language models trained on the entire internet; physical robots have no equivalent repository of real-world data. XDOF builds the collection pipelines, teleoperation tools, and annotation systems that leading AI labs and robotics companies struggle to build themselves — effectively operating as an outsourced data supply chain for the industry.

Project ABC: the largest dataset of its kind

In collaboration with UC Berkeley's research lab, XDOF is preparing to release ABC, which it says is the largest-ever collection of high-quality robotic training data. The dataset combines remote teleoperation with egocentric operators — humans wearing body sensors — performing everyday tasks such as folding clothes and aligning boxes. The company plans to recruit and train data-collection teams worldwide, both teleoperators who control robots remotely and egocentric operators who capture body motion.

Customers and competition

A previous TechCrunch report noted XDOF already works with 20 customers, including several leading AI labs. Competitors include Mecka AI, as well as human-data platforms expanding beyond LLMs — notably Scale AI and Micro1. Deal terms with 8VC are not finalized, the total raise amount has not been disclosed, and it is unclear whether the valuation includes the new capital. XDOF and 8VC did not respond to requests for comment.