Chinese artificial intelligence company Zhipu AI announced on September 13 that it has completed a funding round of approximately $5 billion (about 33.65 billion yuan), according to IThome (IThome). The round comprises about $2 billion (15.68 billion Hong Kong dollars) in share placement and about $3 billion (20.14 billion yuan) in convertible bonds.
The funds will be directed primarily toward the research and development of the next-generation GLM foundation models and a fully self-training (FST) system, as well as the deployment and upgrade of large-scale training, production inference, computing resources and related technology infrastructure.
In its announcement, Zhipu described fully self-training as a process where the next-generation GLM is trained in an environment built by the previous generation, creating a recursive self-improvement loop. Specific investments include automated generation and screening of training data, building task environments, enhancing long-horizon reasoning capabilities, and domestic chip adaptation, operator development and inference optimization.
Dual focus: model capability and computing efficiency
Zhipu stated that this R&D arrangement targets both model capability and computing efficiency. Fully self-training explores having the model itself participate in building training resources, expanding data and task environments for subsequent iterations. Chip adaptation and inference optimization aim to increase effective output under the same hardware conditions. By investing in both paths simultaneously, Zhipu aims to scale training while exploring more efficient ways to improve its models.
The company also launched a city-wide coding initiative in Hangzhou, offering individual users 44% off quarterly cards and 51% off annual cards, though details were not elaborated in the announcement.
This funding round comes amid intense competition in China's AI sector, where companies like Baidu, Alibaba and Tencent are also investing heavily in large language models. Zhipu, often described as one of China's "AI tigers," has been a prominent player in the field, with its GLM series powering various applications.
The announcement did not specify the valuation of the company after the round, nor did it disclose the names of the investors. However, the size of the funding underscores the strategic importance of AI foundation models in China's technology landscape.
