Chinese AI Startup Mosi Intelligent Raises Over RMB 1 Billion in Pre-A Round

On October 12, Mosi Intelligent, a Chinese large-model company focused on context-aware interaction intelligence, announced that it had completed a Pre-A financing round of more than RMB 1 billion. The round was led by Huakong Fund, with participation from Guokai Sci-Tech and Shanghai Science and Technology Innovation Fund. Mosi Intelligent said the proceeds will be used for R&D and capability building in full-modal real-time interaction intelligence, as well as global commercialization.

Founder and chief scientist Qiu Xipeng is a distinguished professor at Fudan University and vice dean of Shanghai Chuangzhi Institute. He began researching pretraining in 2016. In his view, human expression always happens in specific contexts; the timing, tone, gaze, and surroundings of a sentence can change its meaning. What is truly scarce is not a better chat model, but an interaction intelligence that lets AI remain present in the world.

Mosi defines this capability as full-modal real-time interaction intelligence: language, sound, images, environment, and device state continuously enter the same interaction process, and the resulting understanding is converted into speech, content, tool calls, and device actions. Over the past six months, the company has released more than 20 models, including MOSS-TTS and MOSS-Audio, covering speech generation, audio understanding, speech recognition, visual understanding, and content generation, and it continues to open-source its work.