By Yang Jianyong

This year, Zhipu has successfully secured a total of 75.6 billion Hong Kong dollars in funding from the market. This impressive sum includes 4.9 billion Hong Kong dollars from its Initial Public Offering (IPO) in January 2026, 31.4 billion Hong Kong dollars from a private placement in July, and an additional 39.3 billion Hong Kong dollars raised recently through a combination of private placements and bond issuances. Approximately 60% of these funds are earmarked for the research and development of the next-generation GLM foundational model and a fully self-trained system, as well as for upgrading the computing power infrastructure.

Raising 75.6 Billion Hong Kong Dollars to Drive Large Model Innovation

Since the introduction of ChatGPT, artificial intelligence has ushered in the era of large models. Consequently, a plethora of large models have emerged, with over 1,100 generative AI models registered in China alone.

It is noteworthy that as large model technology continues to evolve, large model services have transitioned from initial knowledge-based Q&A systems to Agent AI, which is capable of executing complex tasks.

Faced with the rapid advancement of Agentic AI, the consumption of Tokens has surged, particularly in Agent AI applications, leading to increased Token usage. As a result, model developers still require substantial financial support to meet these demands. Global cloud giants such as Microsoft and Amazon have boosted their AI spending to over 810 billion USD this year.

It is crucial to recognize that the competition in the AI large model sector hinges on the computing power of AI infrastructure, which is a technology- and capital-intensive industry. Consequently, significant computational resources are invested in model training, with increased spending in the AI sector aimed at enhancing model competitiveness.

Of course, these cloud giants possess ample cash reserves and substantial free cash flows. As a leading domestic large model developer, Zhipu, having invested heavily in computational resources for AI model training, has chosen to raise funds through market placements and bond issuances to develop models and meet demand with additional computing power resources.

Zhipu's Revenue Soars Nearly 400%, Pioneering a New Path in Large Model Commercialization

However, the global expansion of AI infrastructure capacity has kept computing costs high, exacerbated by industry competition and other factors. This has directly caused Zhipu's gross margin to plummet from 56.3% in 2024 to 26.38% in the first half of 2026; losses have also continued to widen. In the first half of 2026, adjusted losses reached 1.964 billion yuan, up 12.1%.

Not only Zhipu, but the entire industry finds it exceptionally challenging to profit from large models. Even OpenAI is operating at a loss, with losses reaching 38.5 billion USD last year, while revenue was only 13 billion USD.

AI large model technology is rapidly evolving, and the commercialization process is showing signs of growth. Zhipu primarily relies on Model-as-a-Service (MaaS) and commercialization platforms to provide general-purpose large model services to institutional clients and individual users. In the first half of 2026, the MaaS platform had over 7.4 million users, and the commercialization of AI large models is accelerating, with revenue also showing explosive growth.

In the first half of 2026, Zhipu's revenue reached 954 million yuan, a year-on-year increase of 399.7%. With continuous improvements in GLM's capabilities in Agent planning, tool invocation, sustained execution, and long-term tasks, the business has expanded from single-point Coding applications to high-value scenarios such as software engineering, cybersecurity, data analysis, and complex automation. Its business model has evolved from selling models and API calls to selling subscriptions and end-to-end task results.

This shift in business model has enabled the open platform and API business revenue to reach 825 million yuan in the first half of the year, a 27-fold increase year-on-year, with revenue share rising from 15.2% in the same period last year to 86.5%, becoming a key business for Zhipu. Critically, the rapid growth in revenue scale confirms the enormous potential for the commercialization of AI large model technology.

In Conclusion

In the era of generative AI, the commercialization of large models is on the rise, present everywhere from the cloud to the edge. Large models are reshaping the global technology landscape, with OpenAI and Anthropic seen as global benchmarks, with valuations approaching 2 trillion USD. Zhipu, as the world's first AI large model company to list on the capital market, has also garnered significant market attention, reaching a peak market cap of 1.33 trillion Hong Kong dollars.

Unfortunately, its current market cap stands at only 364 billion Hong Kong dollars, a reduction of nearly 1 trillion Hong Kong dollars from its peak, a 70% decline. Numerous factors have contributed to this market volatility, including the impact of competitor Yuezhi Anmian's powerful Kimi K3 model.

At the same time, in the large model arena, ambitions are high, but reality is tough, with short-term profitability proving elusive, undoubtedly fueling the formation of an AI bubble. Particularly when revenue cannot support a trillion-dollar market cap, this triggers a reassessment of value, prompting the market to reevaluate AI implementation and profit expectations.

Only by steadily promoting the development and commercialization of large models, and creating technological differentiation + commercial sustainability, can long-term value be achieved. The author reiterates the importance of viewing the long-term development potential of AI large models rationally.

The views expressed herein represent personal opinions only. The author focuses on in-depth observations of technology sectors such as AI large models, artificial intelligence, industrial robots, IoT, cloud computing, and smart hardware.