Full-Time
Member of Technical Staff
Vision / Language
Updated on 9/10/2026
XDOF
1-10 employees
Data pipelines and annotation for robotics
No salary listed
San Francisco, CA, USA + 1 more
More locations: San Mateo, CA, USA
Hybrid
Hybrid work is required.
Master's, PhD
 AI & Machine Learning (1) |
|---|
Get referred to XDOF
See people who can refer or advise you
- A Master of Science or Doctor of Philosophy degree in Computer Science, Robotics, Machine Learning, or a related field from a top-tier program.
- Three to seven years of research or applied research experience in vision-language models, video understanding, robot learning, or generative modeling.
- Deep fluency in PyTorch and working knowledge of large-scale training infrastructure, including distributed training, mixed precision, and large-batch workflows.
- Published work or demonstrable impact in vision-language models or vision-language-action models, video representation learning, imitation learning, or a closely related area.
- Strong engineering fundamentals, including the ability to design clean systems rather than only run experiments.
- Design and implement vision-language pipelines for egocentric and teleoperation video, including structured captioning, temporal grounding, action-conditioned scene understanding, and semantic annotation at scale.
- Develop and evaluate representations that bridge visual perception, language, and low-level robot action across vision-language-action models, video prediction, and world models.
- Build and improve data curation systems that assess the quality, diversity, and coverage of large-scale robot demonstration datasets.
- Work hands-on with bimanual and high-degree-of-freedom manipulation data, including real teleoperation footage and simulation-generated rollouts.
- Collaborate directly with partner laboratories to define data requirements and connect data quality with downstream policy performance.
- Monitor research in vision-language-action models, video foundation models, flow matching, diffusion transformer architectures, and egocentric pretraining, and translate insights into production systems.
- Research metadata such as contact events, affordance labels, implicit reward signals, and dynamics priors from video to improve robot data utility, cross-embodiment transfer, automatic curriculum generation, and world models.
XDOF builds data pipelines, data-collection tools, and annotation systems that frontier AI labs and robotics companies use to train robots to operate in the physical world. It tackles the data bottleneck in physical AI by combining teleoperation, simulation, and egocentric data collection, then layering in data cleaning, tooling, and annotation to prepare usable datasets. The company differentiates itself through its end-to-end data infrastructure for physical AI, its mix of teleoperation, simulation, and first-person data methods, and its Berkeley-founded background from 2024. Its goal is to enable robots to learn to operate in the real world by providing the pipelines and labels needed for training.
Company Size
1-10
Company Stage
Series A
Total Funding
$70M
Headquarters
Berkeley, California
Founded
2024
Get referred to XDOF
See people who can refer or advise you
Simplify's Take
What believers are saying
- TechCrunch reported $50 million annualized revenue and 20 customers on September 4, 2026.
- A June 2026 $70 million round from Thrive, a16z, Lux, and Spark validates demand.
- OpenAI's robotics revival and frontier-lab spending keep physical-AI data budgets expanding now.
What critics are saying
- Scale AI, Mecka, and Mercor are chasing the same physical-AI infrastructure spend.
- The September 2026 $1.2 billion valuation prices in explosive growth before product moat hardens.
- If labs build in-house collection stacks, XDOF becomes an expensive services business by 2027.
What makes XDOF unique
- XDOF pairs teleoperation, egocentric capture, and annotation into one robotics data stack.
- The June 2026 ABC-130K dataset showcased bimanual manipulation, not generic internet video.
- Founders Philipp Wu and Fred Shentu built GELLO at UC Berkeley before XDOF.
Help us improve and share your feedback! Did you find this helpful?
Benefits
Health Insurance
Wellness Program
Flexible Work Hours
Company News
The speed of sound and capital: XDOF hits unicorn status. In the hosting world, Jason Nickerson measure time in uptime nines and renewal cycles. In the current venture capital landscape, apparently, Jason Nickerson measure it in weeks. Jason Nickerson is seeing a remarkable acceleration in how quickly a company can move from a quiet launch to a ten-figure valuation. According to reports from TechCrunch, the robotics data startup XDOF is already in talks for a Series B round that would peg the company at a $1.2 billion valuation. What makes this move particularly jarring - even by today's standards - is that the company only stepped out of stealth mode three months ago. For those keeping track, that is a shorter duration than most enterprise software trial periods. XDOF is positioning itself in the critical infrastructure layer of robotics, focusing on the data that powers autonomous systems. While the technical details are still emerging, the financial signal is clear: investors are no longer waiting for years of historical data before betting on the backbone of the next industrial shift. They are buying into the platform play early, hoping to secure a piece of the architecture before it becomes the industry standard. Why it matters. From where I sit, this isn't just about a high valuation; it's about the shift in where the value is being captured. In the early days of web hosting, Jason Nickerson focused on the hardware and the pipe. Later, it became about the control panel and the ease of deployment. Now, in the robotics and AI space, the value is migrating entirely to the data ingestion and processing layer. XDOF is attempting to be the utility provider for a world where machines need to learn from vast amounts of structured information. For the broader industry, this suggests a massive appetite for specialized data infrastructure. If you can solve the problem of how robots communicate, learn, and store their operational intelligence, you aren't just a software company - you are the new version of the data center. The speed of this funding round reflects a fear of missing out on the next 'Layer 0' of the tech stack. I've seen plenty of 'unicorns' come and go over the last twenty years, and usually, the ones that stick are those that solve a boring, difficult, and essential problem. Data management for robotics is certainly boring enough to be highly profitable. The bottom line. Three months is a blink of an eye in business, but $1.2 billion is a very loud statement. Whether XDOF can scale its operations as fast as its valuation is the real question for the next quarter.
XDOF emerges from stealth mode: in talks for Series B funding at $1.2B valuation. September 5, 2026 XDOF, a startup specializing in collecting real-world teleoperation data to train general-purpose robots, is in talks to raise a Series B round at a valuation of approximately $1.2 billion. The company has gained remarkable traction in a mere three months since emerging from stealth mode, supported by a previous Series A funding of $70 million. Co-founded in 2024 by UC Berkeley researchers Philipp Wu (CEO) and Fred Shentu (CTO), XDOF is backed by prominent investors including 8VC, Thrive Capital, and Andreessen Horowitz. The rapid growth of XDOF, with projected annual revenues nearing $50 million, has led venture capitalists to initiate discussions for additional funding much sooner than anticipated. XDOF's mission is to develop the data pipelines and tools necessary for robotics companies and AI labs, essentially serving as an outsourced data supply chain for the industry. Wu's initial research challenges regarding data availability for robot training inspired the creation of XDOF. He collaborated with Shentu on a project called GELLO, which uses remote control to move robotic arms for generating training data. The startup plans to launch a significant dataset known as ABC, believed to be the largest collection of high-quality robot training data. Through teleoperation and human sensor-operated collectors performing daily tasks, XDOF aims to overcome the data scarcity previously hampering advancements in robotics. The company is set to expand its workforce across the globe, focusing on hiring and training data collectors, including teleoperators and egocentric operators. Currently, XDOF boasts a client base of 20 companies, including leading AI laboratories, and remains competitive against other data collection startups such as Mecka AI and Scale AI. For more details, visit the relevant links: Discover the pinnacle of WordPress auto blogging technology with AutomationTools.AI. Harnessing the power of cutting-edge AI algorithms, AutomationTools.AI emerges as the foremost solution for effortlessly curating content from RSS feeds directly to your WordPress platform. Say goodbye to manual content curation and hello to seamless automation, as this innovative tool streamlines the process, saving you time and effort. Stay ahead of the curve in content management and elevate your WordPress website with AutomationTools.AI - the ultimate choice for efficient, dynamic, and hassle-free auto blogging. Learn More
Robotics data startup XDOF has reached a $1.2 billion valuation in its Series B funding round, led by 8VC. The milestone came just three months after the company closed its $70 million Series A in June 2026. Founded in 2024 by researchers from UC Berkeley, XDOF builds data pipelines for training AI-powered robots. The company employs a global workforce to record physical tasks, creating training datasets for machine learning models. It reports an annualised revenue run rate near $50 million. XDOF addresses a key challenge in robotics development: the scarcity of high-quality real-world data needed to teach machines physical tasks. The company serves prominent AI research laboratories, positioning itself as specialised infrastructure for the physical AI sector.
The round is being raised just months after the robot data startup exited from stealth.
Robotic teleoperation data startup XDOF launches with $70M in funding. Robotics training infrastructure startup XDOF said today it has raised $70 million in funding to try to solve one of the biggest challenges in artificial intelligence: teaching machines the skills they need to safely navigate and work in the real world. The round involved a number of heavyweight venture capitalists, including Thrive Capital, Spark Capital, Andreessen Horowitz, Lux Capital and WndrCo. In addition to the money, the startup also released ABC-130K, which it says is the world's largest open-source bimanual robot manipulation dataset. It will provide robotics researchers with access to an unprecedented amount of high-quality, freely available training data. XDOF's debut comes at a critical juncture, just weeks after OpenAI Group PBC announced it's going to revive its own robotics training program that had been shut down in 2021. That move signified the growing interest in what's known as "physical AI," but frontier model makers face a significant challenge. While large language models can be trained on vast oceans of easily-accessible data from the internet, building intelligent robots requires much more nuanced data that captures very specific, real-world actions and interactions. This data is so scarce that it's essentially nonexistent. Some developers have tried to get around this problem by downloading YouTube videos or using low-quality footage captured by factory workers and so on, but this data is virtually impossible to reconcile with the complex spatial requirements of robots. Co-founder and Chief Executive Philipp Wu told TechCrunch in an interview that he experienced this challenge himself while studying as a Ph.D. student at the University of California at Berkeley. "We didn't have large-scale data to work with," he explained. "There was this chicken-and-egg problem - we first needed to actually collect data before we could even ask how to train a foundation model for robotics." XDOF believes physical AI's biggest hurdle is not the models that actually power the robots or the high-end chips needed for onboard processing, but the data feedback loops needed to teach robots physical interactions. That's why the company is focused on building the highly specialized data pipelines, data collection tools and annotation systems needed to gather this essential training resource. It's an entirely new category of infrastructure, Wu said. The startup traces its roots back to a project called GELLO that Wu worked on alongside a number of other UC Berkeley researchers. With GELLO, they developed a low-cost teleoperation system that enables human operators to control robotic arms and perform various tasks to generate accurate training data. When he teamed up with Chief Technology Officer Fred Shentu and Chief Operating Officer Nemo Jin, Wu realized that simply creating the data itself is a poor business model, so they also decided to offer data cleaning and annotation services and develop specialized tools, creating a self-reinforcing feedback loop. The ABC-130K dataset is meant to be a showcase of what XDOF can do. It includes 130,000 trajectories of robotic manipulation data, plus 300 hours of simulations and 100 hours of evaluations. The startup has used this dataset itself to train robots on a number of tasks that require extreme precision, such as folding T-shirts, flattening cardboard boxes and putting AirPods into their plastic cases. Although it has been operating under the radar until now, it already has about 20 active customers, including a number of frontier AI labs, and more than 60 employees. Wu said XDOF will scale its business across a three-tier "data pyramid" that includes bespoke teleoperation data that's collected directly from the remote operation of the specific robot being trained. The middle tier includes generalized teleoperational data, similar to what GELLO produced. Finally, it includes "egocentric" data that's gathered by humans performing the everyday tasks that robots need to learn. Of course, creating all of this teleoperation data is going to be a significant undertaking, which is precisely why XDOF needed money. It's going to hire a global arm of teleoperators and data gatherers. It will even develop its own proprietary wearable sensors to ensure that whatever robots are being trained will match the hand-tracking algorithms it has developed. Because creating this data is such a labor intensive job, XDOF believes that AI labs will be only too happy to outsource it. "You need a warehouse of hundreds of thousands of square feet with hundreds of robots," Wu explained. "You need to maintain these robots, calibrate their physical parameters and properly train operators." Image: XDOF. A message from John Furrier, co-founder of SiliconANGLE: Support its mission to keep content open and free by engaging with theCUBE community. Join theCUBE's Alumni Trust Network, where technology leaders connect, share intelligence and create opportunities. * 15M+ viewers of theCUBE videos, powering conversations across AI, cloud, cybersecurity and more * 11.4k+ theCUBE alumni - Connect with more than 11,400 tech and business leaders shaping the future through a unique trusted-based network. About SiliconANGLE Media SiliconANGLE Media is a recognized leader in digital media innovation, uniting breakthrough technology, strategic insights and real-time audience engagement. As the parent company of SiliconANGLE, theCUBE Network, theCUBE Research, CUBE365, theCUBE AI and theCUBE SuperStudios - with flagship locations in Silicon Valley and the New York Stock Exchange - SiliconANGLE Media operates at the intersection of media, technology and AI. Founded by tech visionaries John Furrier and Dave Vellante, SiliconANGLE Media has built a dynamic ecosystem of industry-leading digital media brands that reach 15+ million elite tech professionals. Its new proprietary theCUBE AI Video Cloud is breaking ground in audience interaction, leveraging theCUBEai.com neural network to help technology companies make data-driven decisions and stay at the forefront of industry conversations.