Zhipu AI, a Chinese artificial intelligence company based in Beijing, completed a $300 million funding round in 2023, securing investment from major technology and financial firms including Alibaba Group, Tencent, Meituan, Ant Group, Xiaomi, and HongShan. The round, which amounted to approximately 2.5 billion yuan, contributed to the company's rapid growth during a period of intense competition in the Generative AI sector. The company later rebranded internationally as Z.ai in 2025 and became China's first major Large language model company to hold an initial public offering, listing on the Hong Kong Stock Exchange in January 2026 with a market capitalization of US$62 billion as of August 2026.
The funding round positioned Zhipu AI among the leading developers of open-weight language models in China, alongside other domestic firms referred to as the country's six "AI tiger" companies by investors. The company's flagship product line, the GLM (General Language Model) family, has competed with models from international developers such as OpenAI and Anthropic. By 2024, the International Data Corporation ranked Zhipu AI as the third-largest LLM market player in China.
Origins and Company Background
Zhipu AI was founded in Beijing in 2019 by Jie Tang, emerging as a spin-out from Tsinghua University. The company's formal Chinese name is 北京智谱华章科技有限公司, often rendered in English as Beijing Zhipu Huazhang Technology Co., Ltd. The term "Zhipu" has occasionally been translated by Chinese researchers as "Wisdom Spectrum." In 2025, the company changed its official name to Knowledge Atlas Technology JSC Ltd.
Researchers associated with the company published a paper at the Association for Computational Linguistics conference in May 2022 introducing the GLM training algorithm. This algorithm employed a strategy termed "autoregressive blank infilling," which constructs cloze tests by randomly removing segments of input text and then trains the model to autoregressively regenerate the removed text. The first GLM version preceded the broader AI boom that captivation around generative models.
The $300 Million Funding Round
During 2023, Zhipu AI raised approximately 2.5 billion yuan - about 350 million US dollars - in a financing round featuring a consortium of prominent investors. Alibaba Group and Tencent led the effort, with participation from Meituan, Ant Group, Xiaomi, and HongShan. The influx of capital enabled the company to expand its research and development of GLM models and infrastructure.
This investment came during a period of intense private funding activity in Chinese AI, as both domestic technology giants and venture capital firms sought to back players in the race for sophisticated language models. The round followed the earlier commercial success of chat-based assistants and highlighted the strategic importance of Chinese companies in global AI development.
Subsequent Financing and Valuation
In May 2024, the Saudi Arabian finance firm Prosperity7 Ventures participated in a US$400 million financing round for Zhipu AI, valuing the company at approximately US$3 billion. This larger round further expanded the company's investor base internationally and underscored its growing role in the global AI ecosystem.
By 2026, Zhipu Ai's valuation had increased substantially following its IPO. The company's success in raising funds and eventually listing on the stock exchange made it a notable case in the intersection of Chinese tech entrepreneurship and government-backed support for innovation.
Model Releases and Technical Development
Throughout the 2020s, Zhipu AI released a series of GLM models with notable technical milestones. In July 2024, the company debuted Ying, a text-to-video model capable of generating a six-second video clip from text prompts in about 30 seconds. The company then released the GLM-4-Plus in August 2024 and the GLM-4-Voice in October 2024, with the latter an end-to-end speech language model able to adjust tone or dialect.
In July 2025, Zhipu AI released the GLM-4.5 and GLM-4.5 Air models, as releasing its models under the free and open-source MIT License, and rebranded its international brand to Z.ai. The GLM-4.5 can function on eight NVIDIA H20 chips. In August 2025, Z.ai announced compatibility of its GLM models with Huawei's Ascend processors. On 11 August 2025, it released the GLM-4.5V, a vision-language model (VLM) with 106 billion parameters.
Further releases came in late September 2025 with the GLM-4.6, which marked the first integration of FP8 and INT4 quantization on Cambricon Technologies chips and also supported native FP8 on Moore Threads GPUs. In December 2025, Z.ai released both the GLM-4.6V and GLM-4.7 models.
In 2026, Z.ai continued its cadence of releases: GLM-5 came on 12 February 2026, followed by GLM-5.1 in late March 2026 (open-sourced on 7 April 2026 with a 10% API price increase). On 16 June 2026, GLM-5.2 introduced a one-million token context window, up from the 200K-window of GLM-5.1. The most recent model release as of August 2026 is GLM-5.3, which retained the same 743 billion parameter base as GLM-5.2 but with improved post-training.
Products and Services
The company offers a suite of products beyond its core GLM models, which have been released under the MIT License since July 2025. In addition to the language models, Z.ai developed the Ying text-to-video model, the AutoGLM AI agent for smartphones that uses voice commands to handle tasks such as ordering items from nearby stores or repeating orders from shopping history, and the ZCode AI coding development environment.
ZCode, released on 2 July 2026 for Windows, macOS, and Linux, includes an AI coding agent that can read and modify project files, execute terminal commands, work with Git, and use a built-in web browser. It supports models from OpenAI, Anthropic, Moonshot AI, and others via OpenRouter.
The company also owns AMiner, a research database created by co-founder Jie Tang in March 2006, similar to Google Scholar. Z.ai maintains offices in the Middle East, United Kingdom, Singapore, and Malaysia, with innovation centers across Southeast Asia as of 2025.
Regulatory and Market Context
In January 2025, the United States Commerce Department placed Zhipu AI (then Zhipu) on its Entity List, citing national security concerns. This designation restricts US-based companies from doing business with the firm without export licenses, adding a layer of complexity to its international operations.
In May 2025, Zhipu disclosed that it had won a 61.28 million yuan government contract for city projects in Hangzhou. The company has benefited from Chinese state policy and investment, while having to handle international sanctions and compute resource limitations.
At the end of February 2026, Z.ai's shares fell by 23% in response to a shortage of compute resources, which led to user complaints and restrictions on new signups, along with public calls for support. Later that spring, the company's share price rebounded, with a 11.5% rise upon the release of GLM-5.1 in April 2026.
Impact and Recognition
Z.ai is considered to be the third largest LLM player in China according to the International Data Corporation, and one of the nation's "AI tigers" for investors as of 2024. Its trajectory, from a Tsinghua University spin-out to a publicly listed company in under a decade, has marked a notable path in the rapidly evolving landscape of Chinese AI developers. Its open-source approach to model development, particularly with the MIT License release, positions it in the broader global trend of Artificial intelligence model accessibility.