Robotics infrastructure startup XDOF is concluding negotiations for a Series B funding round expected to value the company at approximately $1.2 billion, according to venture capital filings and sources familiar with the matter. Led by venture firm 8VC, the prospective deal arrives barely three months after XDOF emerged from stealth, driven by annualized revenues rapidly approaching $50 million.
Unprecedented Growth in Embodied AI Infrastructure
The astonishing valuation surge underscores a dramatic acceleration in institutional demand for physical robotics data. Co-founded in 2024 by University of California, Berkeley researchers Philipp Wu and Fred Shentu, XDOF initially raised a $70 million Series A round in June backed by Thrive Capital, Andreessen Horowitz, Lux Capital, and Spark Capital. Although executives had not intended to solicit fresh capital so quickly, aggressive inbound investor interest forced an expedited timeline.
Industry analysts view XDOF as an indispensable bottleneck solver for embodied artificial intelligence. While large language models trained on trillions of words scraped from the public internet, autonomous physical machinery lacks a comparable digital pre-existing repository. XDOF bridges this fundamental gap by operating an outsourced data-supply chain, building dedicated pipelines, collection tools, and annotation architectures that frontier laboratories cannot economically construct internally.
The Technical Genesis Behind Remote Manipulation
The foundational technology powering XDOF originated during Wu’s doctoral research into how autonomous systems process spatial data. Frustrated by the severe scarcity of high-fidelity robotic demonstration datasets, Wu collaborated with Shentu to invent GELLO, an open design, low-cost teleoperation frame. Their framework allowed human operators to remotely manipulate robotic arm joints with extreme precision, producing an influential research paper that captivated top academic and corporate hardware laboratories.
Translating academic breakthroughs into enterprise scalability, XDOF commercialized the framework into a sophisticated multi-modal data capture network. The startup deploys global teams of remote teleoperators steering physical machinery along with human collectors wearing specialized body sensors. These operators perform routine tasks such as folding garments, opening containers, and flattening boxes, translating real-world tactile movements into pristine algorithmic training inputs for advanced physical neural networks.
To solidify its position as the primary data platform for physical artificial intelligence, XDOF is partnering with the UC Berkeley AI Research lab. Together, they plan to publicly release the ABC dataset, which researchers describe as the largest collection of high-quality robot manipulation data ever assembled. This initiative aims to establish standardized benchmarks across the industry while showcasing the resolution and depth of XDOF’s collection pipeline.
High Revenues and Expanding Enterprise Adoption
XDOF’s commercial traction has scaled with extraordinary velocity over the past two quarters. Enterprise documents confirm the company currently serves at least 20 tier-one clients, including several leading frontier artificial intelligence laboratories developing humanoid platforms. This rapid customer acquisition pushed annualized revenues toward the $50 million threshold, proving that specialized physical data collection carries margins and demand profiles comparable to software data labeling.
Venture capital partners frequently draw comparisons between XDOF and existing data infrastructure giants like Scale AI or Mercor, which catalyzed the early text-based artificial intelligence expansion. However, physical real-world environments introduce complex edge cases, sensor calibration errors, and mechanical latency that traditional software labelers cannot resolve. XDOF’s purpose-built hardware-software integration positions it uniquely to dominate this emerging segment of the robotics technology stack.
Although final terms of the Series B round remain fluid and subject to prospective closing adjustments, participation from 8VC highlights deep conviction among major silicon valley investors. Neither XDOF nor 8VC issued formal statements regarding the pending transaction, but industry observers emphasize that fresh capital will primarily fund international workforce expansion. XDOF plans to recruit hundreds of additional teleoperators and body-sensor data specialists across key geographies.
The Escalating War for Physical Robotics Data
The race to monetize robotic training datasets comes amid an intensifying land grab within embodied artificial intelligence. Emerging competitors such as Mecka AI, along with established human-data labeling platforms like Micro1 and Scale AI, are expanding their footprints into physical space. Yet, early movers capable of delivering continuous, high-frequency kinetic datasets hold a distinct advantage as hardware developers attempt to train general-purpose autonomous agents.
The core challenge for next-generation hardware remains the ability of robots to generalize across unpredictable, unstructured environments like households and industrial warehouses. By leveraging human movement paired with direct teleoperation, XDOF provides the foundational training ground required for neural networks to learn fine motor control. As venture capital pours into hardware developers, the demand for XDOF’s standardized data pipeline is projected to grow exponentially.
Industry experts point out that training general-purpose robots requires orders of magnitude more physical interaction data than currently exists globally. Without specialized third-party data providers handling hardware teleoperation and sensor annotations, individual robotics labs face prohibitive capital costs. XDOF’s outsourced model democratizes access to rich motor datasets, enabling smaller hardware startups to train complex models without investing millions in proprietary data-collection facilities.
Looking ahead, the impending capital infusion positions XDOF to consolidate its dominance across the physical AI infrastructure market. If the deal closes at the reported $1.2 billion valuation, XDOF will become one of the fastest robotics startups to achieve unicorn status in history. The investment underscores a fundamental market pivot from pure algorithmic design toward raw, real-world data acquisition as the primary catalyst for general robotics deployment.
