The edge AI sector has once again seen a major investment. Mingsheng Intelligence announced today that it has completed a new round of financing worth hundreds of millions of yuan, led by Jingguorui, Guoke Investment, CICC Porsche Fund, Mijiu Capital, and Heji Investment. This round of financing will focus on advancing the research and development of efficient large models on the edge and building an ecosystem, further consolidating its leading position in China's edge intelligence field.

image.png

As one of the earliest companies to systematically layout large models on the edge in China, Mingsheng Intelligence has built a complete technical closed loop from theoretical innovation to product implementation. Its core product—MiniCPM "Mingsheng Xiaogangpao" series of edge-side models—has achieved large-scale commercialization in four scenarios: smartphones, PCs, smart homes, and smart cars, becoming a benchmark for domestic edge AI applications.

Comprehensive achievements in the automotive industry, with top automakers rapidly adopting the technology

Mingsheng Intelligence's progress in intelligent cockpit is particularly notable:   

- In April 2025, the MAZDA EZ-60, developed in collaboration with Chang'an Mazda and Wutong Technology, was officially launched, becoming the world's first mass-produced new energy vehicle equipped with an edge-side large model;  

- In September 2025, the Geely Galaxy M9 flagship SUV was released globally, featuring the MiniCPM multimodal model, supporting natural voice interaction, multi-turn context understanding, and personalized service recommendations, creating a "local AI cockpit without internet access";  

- Additionally, the company has reached a deep cooperation with Volkswagen to promote the integration of edge-side AI in high-end vehicles.

Edge-Side Large Models: The Last Mile of AI Implementation

Under the challenges of latency, privacy, and cost faced by cloud-based large models, Mingsheng Intelligence has adhered to the strategy of "model miniaturization, reasoning localization, and experience intelligence," embedding AI capabilities directly into terminal devices. The MiniCPM series can run offline on resource-constrained devices such as smartphones and car systems, achieving millisecond-level response, ensuring user data remains on the device, and reducing reliance on network and cloud services.