Hong Kong-listed artificial intelligence model developers Zhipu and MiniMax surged on the first day their share lock-ups expired, defying expectations that the release of restricted shares would trigger heavy selling and a sharp pullback. Zhipu jumped more than 19%, while MiniMax rose about 17%, as major shareholders signaled they would keep their stakes and helped calm concerns over a flood of new supply entering the market.
The rally marked a notable shift in sentiment toward China’s AI model sector. Traders had widely expected pressure once previously restricted shares became tradable, especially because the combined value of newly unlocked shares was estimated at nearly HK$90 billion, or about US$11.5 billion. Instead, demand absorbed the new supply, and the strong gains suggested that the market was placing more weight on long-term growth prospects than on short-term technical risks.
The move also lifted the broader Hong Kong technology market. The Hang Seng Tech Index climbed nearly 5% in one session, supported by sharp gains across internet, semiconductor, hardware and platform companies. Alibaba rose more than 12%, Hua Hong Semiconductor advanced about 10%, Lenovo gained around 9%, and Semiconductor Manufacturing International, Kuaishou and Tencent also moved higher.
Market participants said the strength reflected a combination of company-specific confidence, improving policy support from Beijing, and a rotation of funds from global semiconductor names into lower-valued Hong Kong technology shares. The rally suggested that traders are once again looking for growth opportunities in Chinese technology companies after a long period of caution toward the sector.
Share lock-up fears fade
The strongest signal came from the behavior of major shareholders in Zhipu and MiniMax. Typically, the expiration of share lock-ups creates concern that early backers, founders or strategic shareholders may sell part of their holdings, adding supply and weighing on prices. In this case, the opposite happened.
Zhipu’s shareholder base includes JSC International, WT Asset Management, Optimas Capital and Lingyun Light Technology. These shareholders indicated plans to continue holding their positions, easing market fears that the lock-up expiration would lead to large-scale selling.
MiniMax received similar support from key backers. Alibaba and MiHoYo reaffirmed confidence in the company late last month, while MiniMax’s founding team voluntarily extended its own sales restriction to 12 months. That is twice the usual six-month lock-up period and was viewed by traders as a strong signal that insiders remain confident in the company’s long-term prospects.
The market reaction was swift. Rather than falling under the weight of newly tradable shares, both stocks climbed sharply. The performance suggested that traders had overestimated the supply risk and underestimated the strength of demand for AI-related names listed in Hong Kong.
The successful absorption of such a large block of newly unlocked shares was also important because it showed that liquidity conditions were stronger than feared. For high-growth technology companies, especially those tied to AI, share lock-up expirations can often create volatility. In this case, the smooth reopening of supply helped reset expectations for the sector.
A reassessment of AI valuations
The gains in Zhipu and MiniMax came as global financial institutions raised or maintained bullish views on the two companies. The ratings helped reinforce the view that the sector is being reassessed based on commercial progress, product development and revenue potential rather than only on concerns about near-term share supply.
JPMorgan raised its price target for Zhipu to HKD 2,000 from HKD 1,800 and kept an “overweight” rating. The bank pointed to the company’s GLM-5.2 model as a factor that could strengthen its commercial opportunities in open-weight AI technology.
Open-weight models have become an important part of the global AI race because they allow developers, enterprises and cloud platforms to customize and deploy models with more flexibility than fully closed systems. For Zhipu, the market is increasingly focused on whether its model capabilities can translate into recurring business demand and wider infrastructure adoption.
JPMorgan also said the market had largely priced in Zhipu’s year-end annual recurring revenue target of US$1 billion. That means further upside may depend on the company’s ability to expand its open model infrastructure and prove that demand can scale across enterprise and developer use cases.
MiniMax also drew attention from major global banks. Goldman Sachs, Bank of America and Citi each issued “buy” ratings on the company, reflecting expectations that its product pipeline and business shift could support future growth.
Goldman Sachs set a price target of HKD 860 per share for MiniMax. The bank highlighted new differential pricing for the DeepSeek V4 product, under which charges double during peak usage periods. That move was interpreted as an attempt to move away from aggressive price competition and toward a more stable margin structure.
For China’s AI model sector, pricing has become a major issue. Several companies have cut prices in an effort to gain users and market share, raising concerns that the industry could face a margin squeeze similar to earlier competition in cloud computing and internet services. MiniMax’s move toward peak-period pricing suggests a more disciplined approach, with the company trying to balance usage growth and profitability.
MiniMax shifts toward enterprise demand
Bank of America set a target of HKD 500 for MiniMax and focused on the company’s changing revenue mix. According to the bank, MiniMax generated about 70% of its revenue from consumer business last year, but it is now shifting toward enterprise and cloud API operations.
That shift is important because enterprise AI demand may offer more predictable revenue than consumer-facing applications. Companies using AI models through cloud APIs often generate repeated usage, especially when the tools are built into workflow systems, customer service platforms, content generation tools or productivity software.
Bank of America also noted that MiniMax’s previous M2.7 model achieved inference profit margins of more than 40%. Inference refers to the process of running an AI model to generate outputs after it has been trained. Since inference costs can be a major burden for AI companies, higher margins in this area suggest that infrastructure efficiency may become a key advantage.
The bank projected that MiniMax could maintain long-term stability if it continues improving infrastructure efficiency. That view is consistent with the broader market focus on whether AI developers can control computing costs while increasing usage.
Citi set a target of HKD 533 per share for MiniMax, implying roughly 54% upside from recent levels. The bank said it expected continued high revenue growth, supported by the launch of a new-generation video model. Video generation is one of the most closely watched areas of generative AI because it requires significant computing power but also offers large commercial potential in advertising, entertainment, gaming and enterprise media production.
Together, the views from major banks added momentum to a rally already strengthened by shareholder discipline. Traders appeared to treat the lock-up expiration not as a trigger for selling, but as a test that the two companies passed.
Broader technology rally gains force
The strength in Zhipu and MiniMax quickly spread across Hong Kong’s technology sector. The Hang Seng Tech Index rose nearly 5%, with gains extending across several major groups.
Alibaba surged more than 12%, benefiting from its role as both a major technology platform and a backer of AI development. Hua Hong Semiconductor gained about 10%, while Lenovo rose roughly 9%. Semiconductor Manufacturing International and Kuaishou rose around 8%, and Tencent at one point added more than 4%.
The breadth of the rally showed that sentiment was not limited to two newly unlocked AI stocks. Instead, traders appeared to be increasing exposure to Hong Kong-listed technology names more generally.
Several forces helped drive the move. Valuations in Hong Kong equities have remained relatively low compared with many global peers, especially after years of pressure on Chinese internet and technology shares. At the same time, some traders have been taking profits from global memory-chip stocks after strong rallies and redirecting capital toward sectors that appear less crowded.
That rotation is important because global semiconductor-related shares, especially those tied to AI hardware, have attracted heavy buying over the past year. As those trades become more crowded, some market participants are seeking opportunities in areas that may benefit from AI adoption but have not yet priced in the same level of optimism.
Hong Kong-listed AI model developers and technology platforms fit that profile. They trade in a market that has lagged many global equity benchmarks, yet they are linked to themes that remain central to global growth expectations, including cloud computing, AI infrastructure, enterprise software and digital consumption.
Policy support helps sentiment
The rally also came against a backdrop of stronger policy signals from Beijing. People’s Bank of China Governor Pan Gongsheng said at a financial forum that China would continue to deepen financial connectivity, support Hong Kong’s market development, strengthen the city’s role as an offshore renminbi center and improve financial stability.
Those comments were followed by concrete measures aimed at increasing market depth and liquidity. The central bank announced plans to raise the annual quota for the Southbound Bond Connect to 800 billion yuan from 500 billion yuan. The Southbound Bond Connect allows mainland institutions to access Hong Kong’s bond market, and a higher quota can support cross-border financial activity.
The People’s Bank of China also said a key renminbi liquidity facility for banks in Hong Kong would be expanded to 500 billion yuan from 200 billion yuan. That measure increases the amount of renminbi liquidity available in the city’s financial system.
For equity traders, these steps matter because liquidity conditions influence risk appetite. A deeper pool of offshore renminbi funding can make Hong Kong more attractive as a financial hub and may help support trading activity across multiple asset classes.
The policy measures also reinforce Hong Kong’s role as a bridge between mainland China and global capital markets. After a difficult period for Hong Kong equities, traders are watching closely for signs that policy support can combine with stronger company fundamentals to produce a more durable recovery.
Market psychology shifts
The reaction to Zhipu and MiniMax suggests a meaningful change in market psychology. Before the lock-up expiration, the dominant concern was supply. Traders focused on the possibility that nearly HK$90 billion in newly tradable shares could pressure prices and drain liquidity from the sector.
After the lock-up expired, the market reached a different conclusion. The absence of large-scale selling, combined with strong public signals from key shareholders, shifted attention back to growth, business models and AI adoption.
This does not remove the risks facing China’s AI model sector. Competition remains intense, computing costs are high, and the path from technical capability to sustained profit is still developing. Pricing pressure also remains a concern, especially as more companies launch models and compete for enterprise customers.
Still, the recent rally shows that traders are willing to reward companies that can present credible growth stories and demonstrate shareholder alignment. In a sector often driven by fast-changing sentiment, the decision by major shareholders and founders to hold their stakes carried significant weight.
The rise in MiniMax also shows how important business mix has become. Traders are increasingly focused on whether AI companies can move beyond consumer hype and build repeatable, revenue-generating enterprise businesses. Cloud API demand, video generation tools and infrastructure efficiency are now central parts of the MiniMax narrative.
For Zhipu, the key question is whether its open-weight model strategy can scale commercially. If the company can expand usage of GLM-5.2 and related infrastructure, it may justify higher valuations. If adoption slows or costs rise too quickly, enthusiasm could fade.
Hong Kong technology enters a new test
The broader rally in Hong Kong technology shares now faces a key test: whether the move can extend beyond a single session and develop into a more lasting recovery. One-day gains can reflect short covering, technical positioning or rapid capital rotation. A stronger trend would require continued earnings progress, policy follow-through and confidence that AI demand can translate into durable revenue.
For now, the sector has gained a powerful boost from the unexpected performance of Zhipu and MiniMax. Their lock-up expirations were expected to challenge market confidence. Instead, they became evidence of demand strength and shareholder conviction.
The gains also highlight how quickly capital can rotate when valuations are low and a credible growth narrative emerges. Hong Kong technology shares have underperformed for an extended period, but that weakness has left some companies trading at levels that appear attractive to traders looking for alternatives to crowded global AI hardware names.
The combination of AI momentum, stable shareholder behavior and stronger liquidity support has changed the tone of the market. Zhipu and MiniMax have become the clearest examples of that shift, with their sharp gains showing that technical risks do not always lead to selling when the underlying story is strong.
Whether the rally continues will depend on execution. AI model companies must prove they can turn product demand into recurring revenue, manage infrastructure costs and avoid destructive price wars. Hong Kong’s technology sector must also show that recent policy support can translate into improved liquidity and sustained market participation.
For now, the message from the market is clear: fears of a post-unlock selloff did not materialize, and traders are again willing to pay for growth in China’s AI sector when confidence, liquidity and valuation support align.
Explore Hong Kong’s AI stock momentum and crypto parallels in our latest analysis—read today’s AI market outlook next.
Disclaimer: The content on this page is provided for general informational purposes only and does not represent the views or financial advice of Toobit. We make no guarantees regarding the accuracy or completeness of this information and shall not be held liable for any errors, omissions, or outcomes resulting from its use. Investing in digital assets involves risk; users should independently evaluate their financial situation and the risks involved. For further details, please consult our Terms of Service and Risk Disclosure.
