XDOF in talks for Series B at $1.2B valuation just months after stealth
XDOF, a real-world robotics data startup, is in late-stage talks for a Series B led by 8VC at a $1.2B valuation, less than three months out of stealth.
Stock photo for illustration only, not from the actual event
- XDOF is negotiating a Series B funding round at a $1.2 billion valuation led by 8VC
- The startup emerged from stealth less than three months ago
- It focuses on building data pipelines and annotation systems for robotics and AI labs
- XDOF is partnering with UC Berkeley to release the ABC robot training dataset
XDOF, a startup specializing in collecting real-world teleoperation data to train general-purpose robots, is in late-stage negotiations to secure a Series B funding round at a valuation of approximately $1.2 billion, led by 8VC, according to sources familiar with the matter. The final capital amount and whether the valuation factors in the new injection remain unconfirmed as deal terms are still subject to change.
The roots of XDOF trace back to co-founder Wu's PhD research into how robots learn from massive datasets, where he faced a severe shortage of large-scale operational data. Partnering with Shentu, they created GELLO, an affordable teleoperation setup enabling human operators to control robotic arms remotely to generate training data, culminating in a highly regarded robotics paper.
That foundational research birthed XDOF, which investors now view as the Scale AI or Mercor equivalent for physical robotics. Unlike large language models that initially trained on the open internet, physical robots lack an equivalent real-world digital repository, making data acquisition the primary bottleneck in building versatile mechanical systems.
Stock photo for illustration only, not from the actual event
To tackle this, XDOF is collaborating with UC Berkeley's AI Research lab to launch what is projected to be the most extensive high-quality robot training dataset to date, named ABC. The company gathers this telemetry by combining remote robot operation with human collectors wearing body sensors to record everyday chores like folding garments and flattening cardboard boxes.
The rapid valuation surge of a physical data startup highlights how critical physical-world data collection has become for the next generation of embodied AI. As frontier labs run dry of synthetic and internet-scraping sources, specialized outsourced data supply chains like XDOF represent the new frontier of venture capital investment.
Looking ahead, XDOF plans to recruit and train global teams of data collectors, spanning remote steering operators and egocentric sensor wearers. The company previously disclosed partnerships with 20 clients, including several prominent AI labs, while facing emerging competition from rivals like Mecka AI and expanding human-data platforms such as Scale AI and Micro1.
Source: TechCrunch
Found something wrong in this article? Report an issue with this article
Comments
Leave a Comment