BitRobot, a robotics data platform, has released 2,000 hours of navigation data to the public and is using Solana to pay contributors who supply real-world robotics data. The company says the move is meant to advance embodied AI, the field focused on machines that can perceive and act in physical environments.
Solana-based rewards for data contributors
BitRobot has integrated Solana into its reward system. Contributors who submit robotics data earn SOL, the cryptocurrency native to the Solana blockchain. The company is leveraging the network to handle payments, though it has not detailed the exact rate or criteria for compensation.
The reward model is designed to encourage more people to share data from real-world robot operations. That data, in turn, helps train AI systems that need to navigate spaces like warehouses, streets, and homes.
Open-source navigation dataset
The open-sourced dataset includes 2,000 hours of navigation data. It covers a range of environments and scenarios, though the company has not specified the exact sources or the types of robots involved. The release is intended to give researchers and developers a large, ready-made set of examples to work with.
By making the data freely available, BitRobot is betting that broader access will speed up progress in embodied AI. The company has not said whether it plans to release additional datasets in the future.
What the data could enable
Navigation data is a core ingredient for training robots to move through the world without constant human guidance. With 2,000 hours of recorded paths, obstacles, and decision points, AI models can learn to handle similar situations on their own.
The release also signals a growing trend of using blockchain-based incentives to gather training data. BitRobot's approach ties cryptocurrency rewards directly to data collection, a model that could attract contributors who might not otherwise participate.
The dataset is now public, and the reward system is in place. How many contributors will join, and what kinds of data they'll bring, is still an open question.