REPORTED: Sina Tech and Yicai say Mosi Intelligence officially announced the round on 12 October. TSN has not seen the company’s own announcement. No valuation was disclosed.
Mosi Intelligence (模思智能), a Chinese AI company incubated from Fudan University professor Qiu Xipeng’s OpenMOSS team, has completed a Pre-A financing round of more than 1 billion yuan, Sina Tech and Yicai reported on Monday 12 October [1][2]. The round was led by Huakong Fund (华控基金) [1][2]. Neither report gave a valuation. (REPORTED)
Who invested
According to both reports [1][2], the main investors include:
- Huakong Fund (华控基金), lead investor
- CDB Sci-Tech (国开科创)
- Shanghai Sci-Tech Innovation Fund (上海科创基金)
- Xinwei Capital (新微资本)
- Xuhui Capital (徐汇资本)
- Zhuiguang Hard Tech VC (追光硬科技创投)
- Lingang Cece Fund (临港策源基金)
Existing shareholders continuing to invest include Oriza Holdings (元禾控股), Shanghai Sci-Tech Group (上海科创集团), part of Shanghai SIIC (上海国投), and Xinglian Capital (星连资本) [1][2]. Both reports end each list with “等” (“and others”), so the lists may not be complete. Duowei Capital (多维资本) is the exclusive financial adviser for this round and on a long-term basis [1][2]. (REPORTED)
What the money is for
The company said the funds will go to research, development and capability building in “full-modal real-time interaction intelligence” (全模态实时交互智能), and to global commercialisation [1][2]. Sina added that the company will also keep betting on “AI for AI”, using AI in data production, filtering and model iteration [1]. (REPORTED, company statement)
Background
Sina describes founder and chief scientist Qiu Xipeng as a distinguished professor at Fudan University and vice-dean of the Shanghai Innovation Institute (上海创智学院) [1]. In February 2023 he led the team that released MOSS, which Sina calls the first ChatGPT-like open-source model in China [1]. Mosi was incubated from the OpenMOSS team at the end of 2024 [1][2].
Since then it has released models including MOSS-TTS, MOSS-Audio, MOSS-Transcribe-Diarize, MOSS-VL and MOVA, and more than 20 models in speech, vision and real-time multimodal AI in half a year of 2026, continuing to open-source them, according to Sina [1]. (REPORTED)
MOSS Live: not yet released
Sina reports that Mosi’s MOSS Live is “about to be released” (即将发布) [1]. It is described as a front-end interaction model that keeps seeing, listening and speaking, working with a back-end agentic model that handles complex reasoning and task execution [1]. The company names glasses, toys, cars, home devices and robots as new entry points for AI [1]. No release date was given. (REPORTED; UNRELEASED)
What is not known
- The valuation and how much each investor put in.
- When MOSS Live will be released.
Sources
- Sina Tech (新浪科技), “模思智能完成超10亿元Pre-A轮融资，押注下一代AI全模态实时交互基座”, 12 October 2026, 09:46 Beijing time. https://finance.sina.com.cn/tech/shenji/2026-10-12/doc-iniuxmtr7267773.shtml
- Yicai (第一财经), Liu Xiaojie, “模思智能完成超10亿元Pre-A轮融资”, 12 October 2026, 09:14 Beijing time. https://www.yicai.com/news/103389154.html
