Axis Robotics
Decentralized robot data collection and training platform
Performances
Comparison
Details
This content is generated by RD AI and is for reference only
Axis Robotics is a decentralized platform for robot data collection and training. It enables global participants to remotely control virtual robots in a browser to complete manipulation tasks, collecting high-quality trajectory data, and uses simulation augmentation to generate diverse training samples for robot manipulation models. Technical mechanisms include MuJoCo simulation and IsaacSim augmentation.
Follow Updates
Follow Lists
People
About Axis Robotics
Axis Robotics is building the data infrastructure layer for Physical AI, focusing on solving the robotics data bottleneck through a "Simulation First" approach. Its core business is a distributed data engine that combines browser-based simulation, crowdsourced human demonstrations, and automated data processing pipelines. The platform allows users to operate virtual robots directly in a web browser, generating diverse trajectories at scale. These trajectories are then filtered, smoothed, and replayed in high-fidelity simulators like NVIDIA Isaac Sim for validation and augmentation, enabling zero-shot sim-to-real transfer.
The company's key differentiator is its three-layer architecture: infrastructure for democratized data collection, a multiplier layer for hybrid augmentation, and a brain layer for policy training. This vertically integrated pipeline transforms raw crowdsourced data into training-ready datasets for foundation models. As of mid-2026, the data engine has accumulated over 300,000 users and 1.8 million trajectories, with a live public dashboard showing real-time growth. Commercial partnerships with robot embodiment companies and model developers have been established, including a notable collaboration with Booster Robotics to combine teleoperation data with simulation scaling for humanoid robots.
The market problem Axis addresses is the high cost and limited diversity of traditional robotic data collection, which relies on physical hardware and lab environments. By shifting data production to simulation and crowdsourcing, Axis claims to reduce costs dramatically while expanding task and scene coverage. Recent developments include the release of a large-scale open dataset for Franka arms, successful deployment of policies trained on crowdsourced simulation data onto physical robots, and the launch of a commercial data-as-a-service offering. The company is positioning itself as the standard-setter for what constitutes pretraining-ready robotic data.
Axis Robotics, founded in 2025, focuses on physical AI data infrastructure. In March 2026, its main product launched on Base Chain, opening an open network for global data contribution. In July 2026, the company completed a $12 million seed round led by Hack VC, with participation from Nomad Capital and others, validating its simulation-first strategy. A key milestone was the "Little Prince's Rose" campaign, where over 10,000 trajectories collected via web-based crowdsourced teleoperation were successfully deployed on a Franka arm for autonomous watering, demonstrating zero-shot sim-to-real transfer. This proved that large-scale, low-cost web data collection can effectively train embodied AI models, establishing a scalable data engine for the industry.
Over the past six months (2026-03-01 to 2026-09-01), Axis Robotics has made significant strides in building its Physical AI data infrastructure. In March 2026, the company announced a $12M seed round led by hack_vc, with participation from Nomad Capital, PiCore Team Ventures, and top angel investors, and revealed plans to launch its main product on the Base blockchain. By June 2026, the platform reached 1 million generated trajectories and 1,600 hours of validated simulation teleoperation data, marking a major scaling milestone. The company also integrated DAgger (Data Aggregation) into its pipeline, enabling a fully closed, bi-directional data flywheel: task setup → data collection → model training → deployment evaluation → targeted task generation → error-correction data collection. This shift moves beyond static datasets toward continuous, feedback-driven model improvement. Looking ahead, Axis Robotics focuses on expanding its data engine across multiple robot embodiments (including bimanual arms and dexterous hands), supporting longer-horizon tasks with state checkpointing, and opening its modular infrastructure to developers and enterprises. The roadmap emphasizes scalable, standardized data production, global community participation, and the gradual decentralization of core interfaces to build a resilient, open network for Physical AI.
Chris Feng (Co-Founder & CEO): Previously COO at Chainbase (2022-2025), Senior Associate at Matrix Partners China (2019-2021), Associate at Bain & Company (2013-2015). Education: BSBA in Finance & Risk Management from Ohio State University.
Christine (Co-Founder & CMO): Previously Project Manager at Nestlé (2017-2018), Business Analyst at Walmart eCommerce (2016-2017), User Growth at Uber (2015-2016). Education: MS in Supply Chain Management from Washington University in St. Louis; BA in Journalism & Economics from Jinan University.