Zhipu Technology is the first to disclose its IPO prospectus, aiming to become the "first global large model stock"?

On December 19, Beijing Zhipu Huazhang Technology Co., Ltd. (hereinafter referred to as "Zhipu") officially disclosed its prospectus after the hearing.

Zhipu has publicly disclosed its prospectus ahead of MiniMax's parent company, Shanghai Xiyu Technology Co., Ltd. (hereinafter referred to as "MiniMax"), and is viewed by many market participants as more likely to become the "first stock of global large models."

MiniMax has been reported to have passed the hearing of the Hong Kong Stock Exchange, but has yet to publish its prospectus.

Since the title of "first stock" is generally determined by the listing date, if Zhipu and MiniMax both list on the Hong Kong Stock Exchange on the same day, it may be difficult to determine who is the first.

With the disclosure of this prospectus, the outside world has gained a deeper understanding of Zhipu's operating conditions.

From 2022 to 2024, Zhipu's revenues were 57.4 million, 125 million, and 312 million yuan, respectively.

Specifically, Zhipu focuses on MaaS (Model as a Service), with revenues divided into two main parts: local and cloud deployment.

Among them, providing clients with exclusive AI models is the main source of revenue, generating 264 million yuan in 2024, accounting for over 80%.

However, local deployment is more like a "one-time project."

According to the revenue recognition method, Zhipu provides localized deployment solutions in a package format, with pricing determined by model type and scale, the amount of computing resources included, and implementation costs. Package prices can be billed as a one-time fee or annually.

Generally speaking, the pricing for individual local deployment projects is relatively high, which is also evidenced by Zhipu's customer structure, where the top five customers contributed 142 million yuan in revenue in 2024, accounting for 45.5%.

In contrast, the revenue generation method for cloud deployment is more "steady."

Zhipu's cloud deployment business is based on contracts billed according to token consumption, with pricing also related to subscription duration, model type and scale, and the amount of computing resources included; however, this part of the business currently accounts for less than 20%.

To support the above two delivery models, Zhipu has built a full-stack product matrix, which includes various parameter scales of models such as edge small models, economical models, and flagship large models with hundreds of billions of parameters, capable of addressing specific customer needs.

At the same time, this matrix provides various functions such as dialogue, general intelligence agents, code generation, image understanding, text-to-image/video, and voice interaction, achieving comprehensive coverage of various model application scenarios.

Currently, Zhipu is still in significant losses, with losses of 144 million, 788 million, and 2.958 billion yuan from 2022 to 2024, totaling 3.89 billion yuan.

The increasing annual losses are mainly eroded by the computing power services used in R&D expenditures.

In 2024, Zhipu's computing power service fees reached 1.552 billion yuan, which is about five times the revenue during the same period.

Given Zhipu's current cash levels, there is indeed some pressure to cope with the losses; as of the end of June 2025, its cash and cash equivalents amounted to 2.552 billion yuan.

Whether Zhipu, which disclosed its prospectus first, can seize the title of "first stock of global large models" is currently under scrutiny