Fudan’s Qiu Xipeng team raises over 1000000000, and Mosi Intelligence advances a Her-style AI assistant
BIBIBI
AT A GLANCE
Mosi Intelligence announced the completion of an over 1000000000 yuan Pre-A round of financing, to be used for full-modality real-time interaction research and development and global commercialization.
Article
Fudan’s Qiu Xipeng team raises over 1000000000, aiming to build a Her-style AI assistant
Beating AI breaking news: Mosi Intelligence, founded by Fudan University Professor Qiu Xipeng, announced the completion of an over 1000000000 yuan Pre-A round of financing, led by Huakong Fund, with participation from Guokai Kechuang, Shanghai Science and Technology Innovation Fund, and others. The company was founded at the end of 2024 and is led by Li Shimin, a student of Qiu Xipeng, as CEO. The funds will be used for research and development of full-modality real-time interaction technology and global commercialization.
Mosi aims to enable AI to understand human emotions and the surrounding environment while understanding what people say. For example, when someone says “It’s a little cold” in a car,AI needs to combine the in-car temperature, the speaker, and past preferences to determine whether the air conditioning needs to be adjusted. It also needs to know when to respond and when to remain silent.
On the technical side, the team has open-sourced the MOSS-Audio and MOSS-VL-Realtime models. The former can analyze emotions in speech and identify different speakers and ambient sounds. The latter can continuously watch video, answer while watching, and revise its answers based on changes in the visuals.
Mosi is currently integrating speech, vision, memory, and tool-calling capabilities into the MOSS Live real-time interaction system, targeting scenarios such as intelligent cockpits and AI hardware. The company says it has begun deliveries to some enterprise customers. Its long-term goal is to create a AI assistant like the one in the film Her, enabling AI to continuously understand the user’s environment.
Original link https://m.theblockbeats.info/flash/371298
Key points
01
Mosi Intelligence announced the completion of an over 1000000000 yuan Pre-A round of financing.
02
The round was led by Huakong Fund, with participation from Guokai Kechuang, Shanghai Science and Technology Innovation Fund, and others.
03
Mosi Intelligence was founded at the end of 2024 and is led by Li Shimin, a student of Qiu Xipeng, as CEO。
04
The funds from this round will be used for research and development of full-modality real-time interaction technology and global commercialization.
05
The team has open-sourced MOSS-Audio and MOSS-VL-Realtime models.
06
Mosi Intelligence is integrating speech, vision, memory, and tool-calling capabilities to develop the MOSS Live real-time interaction system.
07
Mosi Intelligence says it has begun deliveries to some enterprise customers, with application areas including intelligent cockpits and AI hardware.
AI-assisted interpretation
The following is analysis, separate from reported facts. Verify important claims independently.
Mosi Intelligence hopes to develop a AI system capable of simultaneously processing sound, visuals, memory, and tool operations. According to the example in the article, this type of AI must not only understand the user saying “It’s a little cold,” but also combine the in-car temperature, the speaker, and past preferences to determine whether to operate the air conditioning, and decide when to respond or remain silent.
Why it matters to readers
More than 1000000000 yuan in financing will be invested in real-time multimodal interaction research and development and global commercialization; if the relevant system can achieve actual delivery, it could drive AI applications in intelligent cockpits and AI hardware.
For newcomers, it is worth watching whether this type of AI develops beyond simple chat into an assistant that can continuously understand sound, video, and the environment, and execute operations. The financing amount and technical demonstrations do not mean that the product is already mature or generating commercial revenue at scale.
Risks and unknowns
The evidence does not include formal announcements from the company or investors; the financing amount and participating parties lack independent verification.
“Deliveries to some enterprise customers” does not disclose the customer names, number of customers, revenue, or delivery scale.
The evidence does not provide MOSS Live’s official release date, performance data, or status of deployment at scale.
Related Developments
Loading event timeline…
Build something similar to the one in the film “Her”AI Creating such an assistant is a long-term goal, and whether it can be achieved remains uncertain.
Related concepts
Pre-A round of financing
This term is not in the glossary yet. Browse related concepts in the glossary.
