Table of Contents
Axis Robotics has partnered with OpenRoboto, the Bittensor Subnet 80 project building an open competition for robotics models, as the two teams look to connect large-scale robotics data collection with public model evaluation.
The partnership marks Axis' expansion into the Bittensor ecosystem. Under the collaboration, Axis says it will supply more than 3 million multimodal trajectories to OpenRoboto's Open Data Pool and support the project's benchmarking platform through its Data-to-Model Pipeline.
OpenRoboto, which launched on Bittensor mainnet in July, runs an open model-improvement loop in which miners fine-tune the current base robotics model, submit weights, and compete on randomized LIBERO-Pro environments. A challenger only becomes the new base model if it scores better than the current champion.
Bart Hillerich
Axis is a compounding data engine for physical AI, focused on scalable robotics data infrastructure. Its platform emphasizes task generation, simulation-based data collection, mobile egocentric data capture, and data processing tools designed to convert heterogeneous inputs into model-training-ready formats. The Axis network now includes more than 100,000 contributors and is built to collect robotics data across diverse environments and embodiments.
Together, the teams will work to strengthen the data side of OpenRoboto's miner competition. The teams' original announcement said Axis will act as the core data engine behind OpenRoboto by contributing multimodal trajectories to the Open Data Pool.
Axis will also support OpenRoboto's "authoritative benchmarking platform" with its Data-to-Model Pipeline, which the company says will help enable auditable evaluation for data and models across multiple benchmarks.
