Zhipu AI expands further: new technology subsidiary registered in China

Zhipu AI expands further: new technology subsidiary registered in China

Chinese large language model developer Zhipu AI has set up a new subsidiary in the city of Nanchong, Sichuan Province. According to Chinese company registry data, the new unit is named Nanchong Zhipu Huazhang Technology, has a registered capital of 5 million yuan, and will focus on AI infrastructure development and large-scale data processing services. The move points to a wider geographic footprint and preparation for scaling up computing capacity amid growing competition in the artificial intelligence sector.

Zhipu AI's parent company, Beijing Zhipu Huazhang Technology, retains full control of the new entity, which underscores the clarity of its strategic direction. Liu Debin is listed as the founder and legal representative. The registration comes amid intense development of China's AI sector, where regional authorities actively court technology companies with incentives and investment. Nanchong, though not among the country's top IT hubs, is becoming an increasingly attractive location for computing and data centers thanks to lower electricity and real estate costs than in coastal megacities.

The new subsidiary's core business covers development of foundational and application software for AI, as well as data storage and processing services. This structure lets Zhipu AI divide the work: the parent company remains the center of research and development for advanced language models, while the regional unit concentrates on infrastructure and the practical application of the technology. The approach is typical for fast-growing AI startups looking to optimize costs and meet the requirements of local regulators.

Zhipu AI, which created the GLM language model and competes with OpenAI in the Chinese market, faces growing demand for computing resources. Training and deploying modern large language models requires huge amounts of electricity and server hardware. Creating a subsidiary in a province with more favorable conditions for hosting data centers lets the company cut operating costs without compromising service quality. It also brings certain advantages in terms of local support and integration into the regional ecosystem.

The new registration also reflects the fierce competition in China's AI market. ByteDance with its Doubao model, Alibaba with Qwen and other major players are actively investing in infrastructure. Zhipu AI, although it has received funding from leading investors, must demonstrate that it can scale and meet growing demand from enterprise customers. Expanding its regional network should let the company serve customers in different provinces better and reduce data delivery latency.

More broadly, the move shows how Chinese AI companies are adapting to local realities and trying to build sustainable infrastructure for long-term growth. Creating specialized subsidiaries with clearly defined functions is becoming standard practice in this fast-moving sector. New similar initiatives from Zhipu's competitors are likely in the coming months, as the fight for control over computing resources and access to data remains a key factor for success in the AI development race.