xdof is building foundational infrastructure for robotics foundation models. The company's work centres on a core premise: that high-quality training data is the critical bottleneck standing between today's robotics research and general-purpose robots. To address this, xdof develops and operates the full stack of systems required to produce, store, and work with robotics training data at scale.
The company's technical surface area spans several interconnected domains. Its data collection systems are engineered specifically for robotics, complemented by large-scale teleoperation projects that gather real-world data using specialised robotics hardware. Collected data is managed through exabyte-scale data warehouse infrastructure, designed to handle the volumes that robotics model training demands. Alongside this, xdof develops software toolchains and research platforms that allow frontier labs to explore, acquire, and work with massive robotics datasets.
On the robotics side, xdof is developing core technology for universal robot control, including flexible control systems intended to work across robot types. The company positions itself as an infrastructure provider enabling partners - including leading frontier labs - to advance robotics development at scale rather than building every underlying component themselves.
xdof is composed of engineers and roboticists. Its technical domains include robotics, artificial intelligence, robot control, teleoperation, data engineering, and exabyte-scale data infrastructure.