Mifeng Technology, a one-stop physical-AI data service platform, has closed a new several-hundred-million-yuan round led by China Telecom, with Zhangjiang Group following and Sequoia China and Yuanqi Innovation topping up. Founded about half a year ago, the company has now raised multiple rounds from state platforms, top funds and industry players.
Mifeng argues that physical AI needs large, high-quality real interaction data, denser and noisier than language-model data. Its MEgo series of body-less collection products is in mass production and deployed across factories, logistics, supermarkets, homes, hotels, catering, security, elder care and tourism, capturing environment and hand motion at trajectory precision to the millimetre. The MEgo Engine platform automates preprocessing, spatial reconstruction, multimodal labelling and quality checks, lifting efficiency more than tenfold, and each data loop can raise task success rates by 5 to 10 per cent.
Mifeng runs innovation centres in more than 20 Chinese cities and over five overseas nodes, and supplies training data for base models such as LingBot-VLA 2.0 in partnership with Ant Lingbo. Chairman and chief executive Yao Maoqing calls the round not just capital but cloud-network, scenario and ecosystem synergy. With China Telecom’s cloud-network integration and computing scheduling and Zhangjiang’s open scenarios, the three are building a standardised data infrastructure for humanoids, smart manufacturing and intelligent parks.
As physical AI moves from showing off to real work, the bottleneck is no longer models but the data that trains them. Mifeng is betting that whoever runs the platform-level supply of embodied data, not just another model, owns the layer every robot builder must buy from.
Editor’s note: This is an adapted translation of the original LeiPhone report. It has been trimmed and restructured for readability for an international business audience.