top of page

Zhipu AI Surges 132% in Revenue, Setting China’s LLM Market Ablaze

The Chinese artificial intelligence sector has been witnessing a rapid acceleration in both innovation and investment, and few companies exemplify this trend better than Zhipu AI, formally known as Knowledge Atlas Technology Joint Stock Company. Founded in 2019 by researchers from Tsinghua University, Zhipu has positioned itself as one of China’s “AI tigers,” a cohort of startups developing large language models (LLMs) and agent-based AI solutions to rival global leaders such as OpenAI, Anthropic, and Google. In 2025, Zhipu’s growth trajectory has drawn significant global attention, driven by strong revenue gains, strategic technology deployment, and an ambitious global expansion strategy. The company’s first earnings report since its Hong Kong IPO underscores the rapid commercialization of AI in China and provides insights into the operational and market dynamics shaping the industry. Revenue Performance and Market Impact Zhipu reported a remarkable 132 percent year-on-year increase in revenue, reaching 724 million yuan (approximately USD 99 million) in 2025, despite slightly missing analyst expectations of 760 million yuan (Yicai, 2026). This growth reflects robust market demand for AI models, enterprise integration, and platform services. Notably, the company’s annualized recurring revenue (ARR) from its open platform and API offerings surged to 1.7 billion yuan, highlighting the increasing adoption of AI-as-a-Service among small and medium-sized enterprises (SMEs) and developers. Metric 2024 2025 Growth Revenue (CNY) 312M 724M +132% Net Adjusted Loss (CNY) 2.46B 3.18B +29.1% R&D Spend (CNY) 2.2B 3.2B +45% Gross Profit (CNY) 176M 297M +69% Gross Margin 56% 41% -15pp ARR – Open Platform/API (CNY) 28M 1.7B 60x Despite the revenue surge, Zhipu’s net loss widened to 3.18 billion yuan, primarily due to a 45 percent increase in research and development expenditure. The firm’s gross profit rose 69 percent to 297 million yuan, although gross profit margins contracted from 56 percent to 41 percent, influenced by a larger share of lower-margin cloud deployments. Chief Executive Zhang Peng emphasized the strategic necessity of this investment, stating, “Engineering optimizations and platform scaling are essential to maintaining leadership in a market defined by rapid model iteration and computational demand” (Yicai, 2026). GLM-5 Model and AI Agent Innovations Zhipu’s GLM-5 model represents the next-generation architecture for enterprise AI in China. The company claims that the model approaches or exceeds the capabilities of global rivals, including Anthropic’s Claude Opus 4.5 and Google’s Gemini 3 Pro, particularly in coding and agent-based applications. GLM-5’s open-source release allows long-running agent tasks and enhanced coding automation, reflecting the firm’s strategy of positioning China as a leader in LLM-based enterprise solutions. The expansion of AI agents has been a focal point for Zhipu, with revenue from enterprise AI agent deployments soaring 249 percent to 166 million yuan. Over four million SMEs and developers have engaged with Zhipu’s products, spanning 218 countries and regions, showcasing the company’s ability to scale its AI offerings globally while maintaining a strong domestic market footprint. Strategic Technology Deployment and Domestic Chip Integration One of the most notable aspects of Zhipu’s growth strategy is its accelerated deployment of domestic Chinese semiconductor chips. CEO Zhang Peng highlighted that increased adoption of local chips has become critical in response to both rising computing demand and ongoing U.S. export restrictions limiting access to advanced components. This alignment with Beijing’s semiconductor self-reliance policy not only enhances supply chain security but also positions Zhipu as a benchmark for China’s AI sovereignty ambitions. By integrating domestic chips into AI model training and inference operations, Zhipu mitigates geopolitical risk while achieving scalable compute performance across its GLM-5 and related platforms. Market Positioning Among China’s “AI Tigers” Zhipu’s performance is emblematic of a broader cohort of Chinese AI startups, colloquially referred to as “AI tigers,” which are developing large language models and AI agents to compete with global industry leaders. Listed in Hong Kong in January 2026, Zhipu raised $558 million during its IPO, making it the first pure-play AI model company in China to achieve a public listing. The company’s stock has since surged nearly eight-fold from its IPO price, closing at HKD 915 per share, representing a 32 percent gain in a single session (Yicai, 2026). This performance underscores investor confidence in AI-driven enterprises and indicates strong market appetite for high-growth technology companies that combine enterprise AI solutions, platform services, and agent-based intelligence. Zhipu competes domestically with startups like MiniMax, Moonshot AI, and DeepSeek, as well as major technology conglomerates including ByteDance and Alibaba. Its ability to differentiate through advanced model capabilities, API-driven adoption, and enterprise-focused agent offerings has helped establish a leading position within China’s rapidly expanding AI ecosystem. Token Economics and OpenClaw Integration A distinctive aspect of Zhipu’s business model involves the use of tokens as the primary unit of AI processing. Token usage, which reflects computational demand and operating costs, has surged alongside the adoption of OpenClaw, Zhipu’s open-source AI agent platform. Chairman Liu Debing highlighted that price-volume growth for tokens is sustainable, driven by improved model capabilities and shifting enterprise demand: Low-complexity tasks will leverage low-cost or ad-supported token models. High-complexity tasks requiring high reliability will maintain sustained pricing power. This tiered token approach allows enterprises to balance cost and performance effectively while ensuring predictable revenue growth for Zhipu’s AI platform. Global Expansion and Market Penetration Although China remains Zhipu’s primary market, the firm has begun strategic expansion into Southeast Asia, targeting markets where enterprise adoption of AI services is rapidly growing. While it does not yet compete directly with U.S.-based consumer-facing subscription models, international engagement in B2B sectors is increasingly significant. Nine of China’s top ten internet companies are deeply integrated with Zhipu’s GLM models, reflecting the firm’s market credibility and adoption among high-value enterprise clients. This positions Zhipu not only as a domestic leader but also as a key influencer in the global AI ecosystem. Industry Implications and Competitive Analysis Zhipu’s surge in revenue and market attention has several implications for the global AI sector: Acceleration of AI R&D Investment: Rapid revenue growth incentivizes both domestic and global competitors to increase investments in AI model development. Validation of Enterprise AI Models: Zhipu’s success demonstrates that enterprise-focused AI models, particularly those integrating agentic capabilities, can generate substantial ARR and adoption across diverse sectors. Chip Sovereignty as a Competitive Lever: The firm’s domestic chip strategy underscores the increasing importance of supply chain independence in the geopolitically sensitive AI market. Blueprint for Tokenized AI Economy: The tiered token model highlights a path for sustainable monetization of AI computation-intensive services. Experts such as Professor Ming Zhao, a senior AI researcher at Tsinghua University, note: “Zhipu exemplifies how independent AI model developers in China can achieve both technological parity and market scalability with global counterparts, while simultaneously navigating policy and supply chain constraints.” Challenges and Risk Considerations Despite its impressive growth, Zhipu faces multiple challenges: Widening Net Losses: While revenue growth is strong, R&D-driven losses remain significant, posing sustainability questions if profitability timelines extend. Geopolitical Constraints: Inclusion on the U.S. Commerce Department’s Entity List restricts access to advanced semiconductor technology, which could affect model performance scaling. Competitive Pressure: Rapid innovation in China’s AI ecosystem, coupled with international competitors, may compress margins and require continued differentiation. Conclusion: Strategic Growth and AI Leadership Zhipu AI’s first earnings report represents a milestone for China’s AI industry, illustrating the commercialization potential of LLMs and AI agents in enterprise markets. With a 132 percent revenue increase, expansion of GLM-5 capabilities, domestic chip integration, and a growing global footprint, Zhipu is emerging as a barometer for China’s AI ambitions. The company’s approach demonstrates a sophisticated balance between technological innovation, enterprise adoption, and strategic policy alignment, offering insights into how AI-driven firms can scale revenue while navigating geopolitical and market constraints. For investors, enterprise customers, and global AI observers, Zhipu’s trajectory underscores the rising influence of Chinese AI startups and their capacity to rival established international players. For readers seeking further expert insights into AI commercialization, enterprise LLM deployment, and the broader AI technology landscape, the expert team at 1950.ai, under the guidance of Dr. Shahid Masood, provides authoritative analysis, research, and market intelligence. Further Reading / External References CNBC: Zhipu shares surge after revenue doubles in first earnings report — https://www.cnbc.com/2026/04/01/china-ai-tiger-zhipu-shares-surge-revenue-growth-first-earnings-report.html Yicai Global: Zhipu’s stock soars after annual revenue more than doubles — https://www.yicaiglobal.com/news/zhipus-stock-soars-after-chinese-ai-startups-annual-revenue-more-than-doubles TradingView: Zhipu shares jump 30% after debut earnings fuel China AI buzz — https://www.tradingview.com/news/invezz:e38a32e27094b:0-zhipu-shares-jump-30-after-debut-earnings-fuel-china-ai-buzz/

The Chinese artificial intelligence sector has been witnessing a rapid acceleration in both innovation and investment, and few companies exemplify this trend better than Zhipu AI, formally known as Knowledge Atlas Technology Joint Stock Company. Founded in 2019 by researchers from Tsinghua University, Zhipu has positioned itself as one of China’s “AI tigers,” a cohort of startups developing large language models (LLMs) and agent-based AI solutions to rival global leaders such as OpenAI, Anthropic, and Google.


In 2025, Zhipu’s growth trajectory has drawn significant global attention, driven by strong revenue gains, strategic technology deployment, and an ambitious global expansion strategy. The company’s first earnings report since its Hong Kong IPO underscores the rapid commercialization of AI in China and provides insights into the operational and market dynamics shaping the industry.


Revenue Performance and Market Impact

Zhipu reported a remarkable 132 percent year-on-year increase in revenue, reaching 724 million yuan (approximately USD 99 million) in 2025, despite slightly missing analyst expectations of 760 million yuan (Yicai, 2026). This growth reflects robust market demand for AI models, enterprise integration, and platform services. Notably, the company’s annualized recurring revenue (ARR) from its open platform and API offerings surged to 1.7 billion yuan, highlighting the increasing adoption of AI-as-a-Service among small and medium-sized enterprises (SMEs) and developers.

Metric

2024

2025

Growth

Revenue (CNY)

312M

724M

+132%

Net Adjusted Loss (CNY)

2.46B

3.18B

+29.1%

R&D Spend (CNY)

2.2B

3.2B

+45%

Gross Profit (CNY)

176M

297M

+69%

Gross Margin

56%

41%

-15pp

ARR – Open Platform/API (CNY)

28M

1.7B

60x

Despite the revenue surge, Zhipu’s net loss widened to 3.18 billion yuan, primarily due to a 45 percent increase in research and development expenditure. The firm’s gross profit rose 69 percent to 297 million yuan, although gross profit margins contracted from 56 percent to 41 percent, influenced by a larger share of lower-margin cloud deployments. Chief Executive Zhang Peng emphasized the strategic necessity of this investment, stating, “Engineering optimizations and platform scaling are essential to maintaining leadership in a market defined by rapid model iteration and computational demand” (Yicai, 2026).


GLM-5 Model and AI Agent Innovations

Zhipu’s GLM-5 model represents the next-generation architecture for enterprise AI in China. The company claims that the model approaches or exceeds the capabilities of global rivals, including Anthropic’s Claude Opus 4.5 and Google’s Gemini 3 Pro, particularly in coding and agent-based applications. GLM-5’s open-source release allows long-running agent tasks and enhanced coding automation, reflecting the firm’s strategy of positioning China as a leader in LLM-based enterprise solutions.


The expansion of AI agents has been a focal point for Zhipu, with revenue from enterprise AI agent deployments soaring 249 percent to 166 million yuan. Over four million SMEs and developers have engaged with Zhipu’s products, spanning 218 countries and regions, showcasing the company’s ability to scale its AI offerings globally while maintaining a strong domestic market footprint.


Strategic Technology Deployment and Domestic Chip Integration

One of the most notable aspects of Zhipu’s growth strategy is its accelerated deployment of domestic Chinese semiconductor chips. CEO Zhang Peng highlighted that increased adoption of local chips has become critical in response to both rising computing demand and ongoing U.S. export restrictions limiting access to advanced components.


This alignment with Beijing’s semiconductor self-reliance policy not only enhances supply chain security but also positions Zhipu as a benchmark for China’s AI sovereignty ambitions. By integrating domestic chips into AI model training and inference operations, Zhipu mitigates geopolitical risk while achieving scalable compute performance across its GLM-5 and related platforms.


Market Positioning Among China’s “AI Tigers”

Zhipu’s performance is emblematic of a broader cohort of Chinese AI startups, colloquially referred to as “AI tigers,” which are developing large language models and AI agents to compete with global industry leaders. Listed in Hong Kong in January 2026, Zhipu raised $558 million during its IPO, making it the first pure-play AI model company in China to achieve a public listing.


The company’s stock has since surged nearly eight-fold from its IPO price, closing at HKD 915 per share, representing a 32 percent gain in a single session (Yicai, 2026). This performance underscores investor confidence in AI-driven enterprises and indicates strong market appetite for high-growth technology companies that combine enterprise AI solutions, platform services, and agent-based intelligence.


Zhipu competes domestically with startups like MiniMax, Moonshot AI, and DeepSeek, as well as major technology conglomerates including ByteDance and Alibaba. Its ability to differentiate through advanced model capabilities, API-driven adoption, and enterprise-focused agent offerings has helped establish a leading position within China’s rapidly expanding AI ecosystem.


Token Economics and OpenClaw Integration

A distinctive aspect of Zhipu’s business model involves the use of tokens as the primary unit of AI processing. Token usage, which reflects computational demand and operating costs, has surged alongside the adoption of OpenClaw, Zhipu’s open-source AI agent platform.

Chairman Liu Debing highlighted that price-volume growth for tokens is sustainable, driven by improved model capabilities and shifting enterprise demand:

  • Low-complexity tasks will leverage low-cost or ad-supported token models.

  • High-complexity tasks requiring high reliability will maintain sustained pricing power.

This tiered token approach allows enterprises to balance cost and performance effectively while ensuring predictable revenue growth for Zhipu’s AI platform.


Global Expansion and Market Penetration

Although China remains Zhipu’s primary market, the firm has begun strategic expansion into Southeast Asia, targeting markets where enterprise adoption of AI services is rapidly growing. While it does not yet compete directly with U.S.-based consumer-facing subscription models, international engagement in B2B sectors is increasingly significant.

Nine of China’s top ten internet companies are deeply integrated with Zhipu’s GLM models, reflecting the firm’s market credibility and adoption among high-value enterprise clients. This positions Zhipu not only as a domestic leader but also as a key influencer in the global AI ecosystem.


Industry Implications and Competitive Analysis

Zhipu’s surge in revenue and market attention has several implications for the global AI sector:

  1. Acceleration of AI R&D Investment: Rapid revenue growth incentivizes both domestic and global competitors to increase investments in AI model development.

  2. Validation of Enterprise AI Models: Zhipu’s success demonstrates that enterprise-focused AI models, particularly those integrating agentic capabilities, can generate substantial ARR and adoption across diverse sectors.

  3. Chip Sovereignty as a Competitive Lever: The firm’s domestic chip strategy underscores the increasing importance of supply chain independence in the geopolitically sensitive AI market.

  4. Blueprint for Tokenized AI Economy: The tiered token model highlights a path for sustainable monetization of AI computation-intensive services.

Professor Ming Zhao, a senior AI researcher at Tsinghua University, note: “Zhipu exemplifies how independent AI model developers in China can achieve both technological parity and market scalability with global counterparts, while simultaneously navigating policy and supply chain constraints.”


Challenges and Risk Considerations

Despite its impressive growth, Zhipu faces multiple challenges:

  • Widening Net Losses: While revenue growth is strong, R&D-driven losses remain significant, posing sustainability questions if profitability timelines extend.

  • Geopolitical Constraints: Inclusion on the U.S. Commerce Department’s Entity List restricts access to advanced semiconductor technology, which could affect model performance scaling.

  • Competitive Pressure: Rapid innovation in China’s AI ecosystem, coupled with international competitors, may compress margins and require continued differentiation.


Strategic Growth and AI Leadership

Zhipu AI’s first earnings report represents a milestone for China’s AI industry, illustrating the commercialization potential of LLMs and AI agents in enterprise markets. With a 132 percent revenue increase, expansion of GLM-5 capabilities, domestic chip integration, and a growing global footprint, Zhipu is emerging as a barometer for China’s AI ambitions.


The company’s approach demonstrates a sophisticated balance between technological innovation, enterprise adoption, and strategic policy alignment, offering insights into how AI-driven firms can scale revenue while navigating geopolitical and market constraints. For investors, enterprise customers, and global AI observers, Zhipu’s trajectory underscores the rising influence of Chinese AI startups and their capacity to rival established international players.


For readers seeking further expert insights into AI commercialization, enterprise LLM deployment, and the broader AI technology landscape, the expert team at 1950.ai, under the guidance of Dr. Shahid Masood, provides authoritative analysis, research, and market intelligence.


Further Reading / External References

bottom of page