Anima Technology Pte. Ltd.

Axis Robotics (styled AXIS) is a physical-AI data infrastructure company founded in 2025 and operated by Anima Technology Pte. Ltd., the entity named in the copyright line of its website. Chris Feng is co-founder and CEO. The company generates robot training data by letting a distributed community of contributors complete manipulation tasks in simulation from an ordinary web browser, without physical robots or special hardware, and then cleans and converts those demonstrations into datasets for robot models.

The platform, which the company calls a compounding data engine, has four parts. A Task Generation Engine procedurally creates tasks and environments, randomising objects, layouts and robot embodiments. The Sim Data Collection Platform is an online, human-in-the-loop simulator running on MetaSim, which the company describes as MuJoCo compiled to WebAssembly so that no download is needed; users access it through hub.axisrobotics.ai. A Mobile Egocentric App records real human hand and body motion with a phone. A data-processing pipeline turns raw inputs into model-ready formats for large language models, world models and vision-language-action (VLA) policies. Solutions marketed on the site are embodiment integration, robotics data for model training and industrial sim-to-real transfer. The team has published a technical report and an arXiv paper titled "AXIS: A Growable Community-Driven Data Engine for Scalable Robot Manipulation".

Axis reports concrete collection campaigns. In the "Little Prince's Rose" campaign in January 2026, contributors produced more than 10,000 valid trajectories in three days, and a policy trained on that community data was deployed on a physical robot to water a plant autonomously. A beta in February 2026 drew 18,000 participants who produced nearly 100,000 trajectories across 27 task types in five days; the company compares this with roughly 100 to 300 demonstrations per operator per day for conventional teleoperation. By mid-2026 it claimed more than 100,000 active contributors and, per month, over 1,200 hours of simulation data and more than 20,000 hours of real-world data. It has released Sim Dataset V1, which it says raised success rates on the LIBERO-Plus benchmark when used for pretraining, and announced Sim Dataset V2 and a DAgger dataset for later in 2026. Named partners and customers include Booster Robotics, Manycore Tech, Dexmal, Lotus and Geely Auto.

In July 2026 Axis Robotics announced a USD 12 million seed round led by Hack VC with participation from Nomad Capital, Pi Network Ventures, 10K Ventures and angel investors. Community channels are X, Discord and GitHub; no email contact is published.