ZHIPU Surges Past HK$1,900, Nearly Doubling from June Lows: Where Is ZHIPU’s Ceiling?

Markets
更新済み: 2026/06/18 03:28

ZHIPU’s latest surge isn’t just an isolated price movement—it’s the result of multiple short-term catalysts converging. As of June 18, 2026, according to Gate market data, ZHIPU is trading at HKD 1,905, up 15% since the opening today. The price has nearly doubled from its mid-June low and is now just shy of its all-time high.

Looking at the timeline, mid-June marks a key turning point for this rally. On June 11, Gate officially launched its Hong Kong stock trading service, supporting over 1,000 Hong Kong-listed stocks. ZHIPU quickly became one of the most watched assets on the platform. On June 13, Zhipu announced the full release of its flagship model, GLM-5.2, offering a 1M context window and officially open-sourcing model weights under the MIT license. On June 15, JPMorgan raised Zhipu’s target price to HKD 1,400 (about USD 180), maintaining an "overweight" rating. The stock soared as much as 48% that day. These events unfolded in less than a week, forming the core drivers behind this price rally.

How Model Upgrades Impact Market Valuation

The full release of GLM-5.2 marks a major milestone for Zhipu’s product lineup. This model leads domestically in long-range programming and agent tasks. Early developer reviews suggest its coding capabilities are approaching those of Claude Opus. From a market valuation perspective, improvements in model capabilities directly affect two areas: increased API call volume expectations and stronger customer willingness to pay.

Previously, Zhipu demonstrated pricing power—API prices rose 83%, yet call volume jumped 400%. The full release of GLM-5.2 reinforces this dynamic: stronger models drive higher customer stickiness and deepen competitive moats. The market premium assigned to ZHIPU essentially prices in the assumption that its "model capabilities remain ahead of the curve."

Has Valuation Detached from Fundamentals?

This is the central debate surrounding the current rally. According to Zhipu’s 2025 earnings report, annual revenue reached RMB 724 million, up 131.9% year-over-year, ranking first among independent domestic large model vendors. However, rapid growth comes with steep losses—net loss for 2025 was RMB 4.718 billion, up 59.5% year-over-year, with R&D expenses at RMB 3.18 billion, roughly four times annual revenue.

At current prices, ZHIPU’s price-to-sales ratio (PS) is extremely high. Some analysts estimate its PS at 773x, making it one of the most expensive AI stocks globally. For comparison, competitor MiniMax generated RMB 570 million in revenue but has a market cap less than one-tenth of Zhipu’s. This valuation divergence reflects differing market views on the business models and growth trajectories of the two companies, but it also means ZHIPU’s valuation is heavily driven by expectations—any results falling short could trigger a re-pricing.

Capital Structure and Market Participant Behavior

On June 8, Zhipu was officially added to the Hang Seng Tech Index and the Stock Connect program, with southbound capital net purchases hitting RMB 920 million in a single day. Inclusion in Stock Connect allows mainland investors easier access, directly expanding buying power. From a behavioral finance perspective, index inclusion often prompts passive funds to allocate, while active funds may position ahead of anticipated inclusion.

Additionally, Zhipu is pursuing a secondary listing on the STAR Market, aiming to raise RMB 15 billion. If successful, it would become the first "A+H" large model company. The high valuation premium for AI stocks in the A-share market could further boost sentiment. However, it’s important to note that speculation around secondary listings often carries "buy the rumor, sell the news" risk—positive developments may trigger profit-taking once realized.

The Underlying Logic of Bull-Bear Divergence

Market views on ZHIPU are sharply divided. Bullish arguments rest on three pillars: world-class model capabilities (GLM-5.1 Coding surpassing GPT-5.4), projected 2026 revenue growth of 316% to RMB 3 billion, and dual capital catalysts from Stock Connect and A-share listing.

Bearish arguments focus on valuation bubbles and competitive dynamics. On June 17, Hedgeye analysts issued ZHIPU’s first sell recommendation, suggesting a fair short target of HKD 407. They argue the company lacks pricing power, especially amid the AI price war triggered by DeepSeek V4’s launch. Additionally, Ray Dalio has publicly highlighted the risks of an AI bubble. Historically, after hitting all-time highs, ZHIPU’s pullback has reached 35%, and after positive news, the stock dropped 13%, confirming the "buy the rumor, sell the news" market pattern.

Technical Significance and Market Psychology of All-Time Highs

ZHIPU’s 52-week high was HKD 1,993, reached intraday on May 29. The current price of HKD 1,905 (per Gate market data) is just a step away. From a technical analysis perspective, all-time highs often serve as major psychological resistance—these are zones where profit-taking concentrates and new investors weigh whether to chase the rally.

However, technical levels must be interpreted alongside fundamentals. The late-May peak coincided with GLM-5.1’s launch and IPO enthusiasm, while the current price reflects a combination of GLM-5.2’s release, Stock Connect inclusion, and JPMorgan’s upgraded target. The density and quality of these catalysts exceed those seen previously, providing stronger fundamental support for a breakout. Still, as the price nears the previous high, the bull-bear battle intensifies—whether a breakout is confirmed depends on whether new buying can absorb the overhead supply at the high.

ZHIPU’s Evolving Industry Position

Zhipu is transitioning from "China’s leading large model stock" to "a key global AI competitor." As JPMorgan raised Zhipu’s target price, it downgraded competitor MiniMax to neutral. This split in ratings reflects Wall Street’s differing views on the two companies’ strategic paths: Zhipu focuses on large language models, while MiniMax emphasizes a broader product suite including video generation and consumer applications.

From a global perspective, tightening US controls on AI technology exports to China are ironically creating market opportunities for Chinese AI firms. China has made technological self-sufficiency a top priority in its new five-year plan (2026-2030). As a national team company backed by Tsinghua University’s technology transfer, Zhipu enjoys natural advantages under these policy tailwinds. At the same time, AI price wars launched by competitors like DeepSeek are squeezing industry-wide profit margins.

Logical Analysis of Risk Factors

ZHIPU’s lofty valuation embeds at least three layers of risk. First, performance risk—if the projected 316% revenue growth for 2026 fails to materialize, the valuation faces systemic downward adjustment. Second, competitive risk—AI price wars could erode Zhipu’s pricing power and gross margins. Third, sentiment reversal risk—the current valuation fully reflects multiple positive expectations; any changes in macro conditions or industry policy could trigger sharp volatility in this high-beta AI sector.

From a financial perspective, Zhipu’s consolidated gross margin fell from 56.3% in 2024 to 41% in 2025. For every RMB 1 earned, the company burns RMB 6.5, meaning it must rely on external financing to sustain operations. As of year-end 2025, cash on hand was RMB 2.259 billion, with a debt-to-asset ratio of 15.2%—while debt is light, the burn rate makes the capital chain highly dependent on market confidence.

Conclusion

ZHIPU’s breakthrough above HKD 1,900, currently quoted at HKD 1,905, is the result of three converging factors: product iteration (full release of GLM-5.2), capital structure (Stock Connect inclusion), and institutional endorsement (JPMorgan’s upgraded target price). This rally has nearly doubled the price from mid-June lows, putting it just short of its all-time high.

However, the extremely low return on investment (burning RMB 6.5 for every RMB 1 earned), declining gross margin (41%), and a price-to-sales ratio in the hundreds place significant pressure on valuation. Bulls and bears are locked in a debate over whether "high growth can continue to justify high valuation"—bulls believe leading model capabilities and market expansion support the premium, while bears argue the valuation has detached from fundamentals.

As the price approaches its all-time high, the market is re-pricing Zhipu’s value. Whether this pricing is sustainable ultimately depends on the interplay of product commercialization, competitive dynamics, and capital flows.

FAQ

Q1: What are the main drivers behind ZHIPU’s latest rally?

The rally is driven by multiple converging factors: the full release of the GLM-5.2 model has boosted expectations for product competitiveness; inclusion in Stock Connect on June 8 brought in southbound capital flows; and JPMorgan’s June 15 target price upgrade to HKD 1,400 further lifted market sentiment.

Q2: What is ZHIPU’s current valuation level?

At current prices, ZHIPU’s price-to-sales ratio is extremely high, with some analysts estimating a PS of 773x. In 2025, the company reported RMB 724 million in revenue but a net loss of RMB 4.718 billion, reflecting both rapid growth and steep losses.

Q3: How close is ZHIPU to its all-time high?

ZHIPU’s 52-week high was HKD 1,993, reached intraday on May 29. As of June 18, 2026, according to Gate market data, ZHIPU is quoted at HKD 1,905—just a step away from its all-time high.

Q4: What risks should investors watch when considering ZHIPU?

Key risks include: performance risk (can the high growth expected for 2026 be achieved?), competitive risk (AI price wars may erode margins), and sentiment reversal risk (high-valuation sectors are sensitive to macro changes).

Q5: What does the valuation gap between ZHIPU and its competitors mean?

MiniMax generated RMB 570 million in revenue but has a market cap less than one-tenth of Zhipu’s. This divergence reflects differing market views on the two companies’ strategic paths—Zhipu focuses on large language models, while MiniMax covers broader areas such as video generation and consumer applications.

The content herein does not constitute any offer, solicitation, or recommendation. You should always seek independent professional advice before making any investment decisions. Please note that Gate may restrict or prohibit the use of all or a portion of the Services from Restricted Locations. For more information, please read the User Agreement
コンテンツに「いいね」する