The company has attracted backing from investors including Ant Group, Didi, Delian Capital, Shunwei Capital and Baidu Ventures across its recent rounds. Earlier this year, Simple AI disclosed that it had completed multiple financings totaling several hundred million yuan.
Founded in 2025, Simple AI is building around one of the main bottlenecks in robot learning: access to large amounts of high-quality manipulation data. Its approach combines human-generated training data, robot learning models and systems designed to transfer learned behaviors onto physical robots.
In August, the company released HiFi-UMI, a robot-free data collection system designed to capture manipulation demonstrations without requiring a physical robot during collection. Alongside it, Simple AI open-sourced HiFi-UMI-2K, a dataset containing 2,000 hours of real-world manipulation data.
The dataset spans large-scale human demonstrations that can be used to train robot manipulation policies, with Simple AI showing that policies trained on the data can then be deployed on physical robots. HiFi-UMI-2K has already recorded more than 160,000 downloads on Hugging Face in the past month.
Simple AI is ultimately targeting general-purpose robots capable of operating across less structured environments, including service and household settings. The company is using its data infrastructure to shorten the loop between collecting human demonstrations, training models and deploying new behaviors onto robots.
The new capital gives Simple AI additional resources to expand that stack as competition intensifies around the data needed to train general-purpose robot intelligence.
