This article breaks down how Zhipu, an AI company listed on the Hong Kong stock exchange, saw its share price surge more than 20-fold in six months. Key takeaways are as follows:

1. Zhipu’s three major share price rallies aligned perfectly with three iterations of its GLM large language model (LLM), showing that capital markets now anchor valuations of AI LLM companies to technical signals from model iterations, rather than quarterly financial reports.

2. Zhipu’s latest GLM-5.2 model now ranks among the top three in global general AI capability rankings, fourth globally for coding ability, and supports a 1 million-token lossless context window while being compatible with all major domestic Chinese AI computing platforms. Launched at a time when top overseas models have been cut off from the Chinese market due to policy restrictions, the model has gained widespread global attention.

3. Beyond technical breakthroughs, four additional factors drove the share price rally: reshaped market perception, amplified discussions from global AI opinion leaders, an extremely low free-float share ratio, and rising API sales volume and pricing. However, Zhipu still faces challenges including an annual net loss exceeding 4.7 billion yuan, an upcoming large-scale share unlock, and insufficient commercialization. Its long-term value will ultimately need to be validated by financial performance.

This article shares valuable insights for brand building, marketing, and industry trend positioning in the AI sector. Key takeaways are as follows:

1. For brand marketing, Zhipu’s timing-aligned marketing strategy is worth learning from. After the U.S. government ordered top overseas models to cut off global services, Zhipu quickly launched an open-source LLM with no geographic restrictions and permissive commercial licensing, establishing a brand positioning of cutting-edge AI openness and sharing. This helped it quickly win recognition from global developers and capital markets.

2. Industry perception of LLMs has shifted, with core brand competitiveness now anchored to technical iteration capabilities. One hard technical breakthrough can completely reshape market perception of a brand and unlock new valuation upside: GLM-5.2’s top-three global ranking directly elevated Zhipu from a top domestic open-source model provider to a member of the global first tier.

3. In terms of industry trends, market demand for open-source LLMs compatible with domestic Chinese computing power is booming. After Zhipu raised its API prices by 83%, API call volume still grew 400% year-over-year. This simultaneous growth in volume and price confirms the huge market opportunity under the broader trend of domestic substitution, and aligning with China’s independent and controllable technology strategy can bring additional strategic value to a brand.

This article outlines market opportunities and potential risks in the LLM industry for industry practitioners. Key insights are as follows:

1. On the opportunity side: policy-driven supply disruptions of top closed-source overseas models have created significant market space for domestic Chinese open-source LLMs. There is currently extremely strong market demand for unrestricted, commercially permissive LLMs that are fully compatible with domestic Chinese computing power, and demand growth has already been validated by commercial data.

2. Key risks to watch: The LLM industry is still in the high-investment growth stage. Zhipu currently has a price-to-sales ratio above 1,000x, with an annual net loss of 4.718 billion yuan that continues to widen. Its commercialization path has not yet been proven, and there is major market debate over whether its valuation follows a new sustainable paradigm or is simply a speculative bubble. A large-scale share unlock is scheduled for July, which will expand the free-float share by more than 200%. The share price, inflated by the previously tiny free float, could see sharp downside correction going forward.

3. A actionable growth path to reference: Focusing on continuous technical iteration and aligning with China’s domestic computing power self-reliance strategy can quickly win capital and market attention and open up substantial growth space.

This article shares key insights for factories looking to capture AI industry opportunities and advance digital transformation. Key takeaways are as follows:

1. Current LLM technology is already mature enough to support factories’ digital transformation needs. The latest GLM-5.2 supports long continuous operation, and can independently complete the full workflow from development and debugging to deployment. It can help factories digitize product design and production management processes, and lower technical development barriers.

2. In terms of commercial opportunity, domestic computing power self-reliance is a core strategic direction for China’s AI industrial chain. GLM-5.2 is already compatible with almost all major domestic AI computing platforms including Huawei Ascend and Cambricon. Factories now have access to mature, usable domestic AI solutions that are not subject to overseas supply restrictions, eliminating the risk of service cutoffs.

3. A key lesson for factory digital transformation: AI technology is advancing rapidly, and the commercial usability of open-source LLMs has improved dramatically. Factories can leverage mature domestic open-source LLM solutions to advance their digital transformation, lower trial-and-error costs, and deploy customized intelligent applications faster.

This article outlines the latest industry trends, customer pain points, and actionable paths for the LLM services industry. Key insights are as follows:

1. Current industry trends: The LLM industry has moved beyond concept popularization to enter a phase of technical competition. Open-source LLMs are gaining growing market recognition and strategic value, and capital markets now treat technical iteration capability as the core valuation anchor for LLM companies. For open-source models, core value lies in ecosystem positioning and technical influence, rather than short-term revenue.

2. Core customer pain points today: First, top closed-source overseas models face major policy uncertainty, with the risk of service cutoff at any time, making enterprise clients hesitant to adopt them at scale. Second, most clients are seeking LLM solutions compatible with domestic Chinese computing power, and there has been a lack of mature top-tier options to meet this need.

3. A proven solution path to reference: Focusing on open-source, commercially permissive domestic LLMs, completing compatibility with all major domestic computing platforms in advance, and launching services during the window of overseas model supply disruption has been validated as a path to quickly win market acceptance. Zhipu’s simultaneous growth in sales volume and pricing provides clear evidence of market demand for this approach.

This article shares valuable reference for AI platform operators on aligning with market demand, optimizing operations, and mitigating risks. Key insights are as follows:

1. The core market demand for AI platforms has shifted toward providing stable, open, policy-safe LLM services compatible with domestic Chinese computing power. Uncertainty around overseas closed-source models has pushed large numbers of developers and enterprise clients to turn to domestic open-source LLM platforms, creating new growth opportunities for domestic platforms.

2. Actionable operational directions to reference: Technical iteration is currently the core growth engine of the LLM industry. Platforms can build influence around LLM technical iterations to attract developers and enterprise users; every top-tier model release from Zhipu has driven sharp growth in commercial metrics and attention, validating the feasibility of this technology-driven growth path. Platforms can also build investment attraction advantages around domestic LLMs and domestic computing power compatibility to attract relevant companies to settle on the platform.

3. Risks to avoid: Valuations across the LLM industry are currently elevated, and most companies are still in the high-investment, loss-making stage. When developing related businesses or recruiting new tenants, platforms need to pay close attention to companies’ commercialization progress to avoid risks from valuation bubbles, and also watch for volatility from capital-side events such as large-scale share unlocks.

This article outlines the latest industry developments in China’s domestic open-source LLM space, and offers high value for research on new shifts in the global AI industry. Key takeaways are as follows:

1. New industry developments: Domestic Chinese open-source LLMs have achieved substantive technical breakthroughs. Zhipu’s GLM-5.2 is the first Chinese model to rank among the top three in global general LLM capability rankings, fourth globally for coding ability, and second globally for agent capability. It is compatible with almost all major domestic AI computing platforms, and has successfully filled the market gap left by policy-driven disruptions to top overseas models, proving that Chinese open-source LLMs have now joined the global first tier.

2. New industry research questions: The valuation logic for LLM companies is fundamentally different from that of traditional enterprises. Zhipu carries a valuation of one trillion Hong Kong dollars against just 724 million yuan in annual revenue and a net loss of nearly 4.7 billion yuan. There is major market debate over whether its valuation follows a sustainable new paradigm or is a capital-driven bubble. Low free-float combined with social media amplification drastically increases share price volatility, and the uncertainty created by large-scale share unlocks also creates new research topics for the industry.

3. In terms of business model, open-source LLMs follow a logic similar to the Android operating system, where core value lies in ecosystem positioning and technical influence rather than short-term revenue, and commercial returns are generated long-term by enabling the entire industrial chain. This offers a completely new business model framework for the AI industry that merits further in-depth research.