Tech
The Dirty Work of Robot Training: XDOF Emerges to Fill the Data Gap
AI Summary
XDOF, a new startup, is addressing the bottleneck in robot training data by building data pipelines, collection tools, and annotation systems for frontier labs and robotics companies. The company has raised $70 million and is already working with 20 customers, including several frontier AI labs. XDOF aims to provide high-quality robot training data to enable robots to interact with the physical world.
The Emergence of XDOF
The race to teach machines to operate in the physical world has led to a new kind of infrastructure business. XDOF, emerging from stealth, is betting that the next great bottleneck in AI isn’t models or chips, but the data feedback loop needed to teach robots how to interact with the physical world.The Data Gap in Robotics
Unlike LLMs that were trained on a vast sea of publicly available text, robots need data that captures physical interaction, and that kind of data barely exists. YouTube videos and footage captured by gig workers are low-fidelity and hard to reconcile with the physical world.Building the Data Pipelines
XDOF aims to build the data pipelines, collection tools, and annotation systems that frontier labs and robotics companies can’t easily build themselves. The company has raised $70 million from Thrive Capital, Spark Capital, a16z, Lux, and WndrCo.The Data Ecosystem
- XDOF has about 60 employees and is already working with 20 customers, including several frontier AI labs.
- The company is partnering with UC Berkeley’s AI Research lab to release the largest collection of high-quality robot training data ever assembled, dubbed ABC.
- ABC includes 130,000 trajectories of robot manipulation data, 300 hours of simulation, and 100 hours of evaluations.