Zhipu AI reported a 400% surge in first-half revenue as demand for Chinese artificial intelligence accelerates. Yet heavy R&D spending and continued losses are shifting investor attention from AI growth towards the increasingly important question of profitability.
Chinese artificial intelligence developer Zhipu AI has reported a 400% surge in first-half revenue, providing fresh evidence of rapidly growing demand for home-grown AI technology. Yet heavy research spending and continued losses are raising an increasingly important question for investors: when will China’s AI boom start producing sustainable profits?
Zhipu delivers extraordinary revenue growth
China’s artificial intelligence investment story gained another significant data point on Monday, 31 August 2026, after Zhipu AI reported first-half revenue of 953.9 million yuan, representing growth of approximately 400% compared with the same period one year earlier.
The Beijing-based company, which became the first major Chinese large-language-model developer to list in Hong Kong, has quickly become one of the most closely watched pure-play AI stocks in the region and its latest figures help explain why.
Demand for artificial intelligence software and services is expanding rapidly across China as companies integrate generative AI into coding, cybersecurity, customer services and business processes. Zhipu has positioned itself directly within that transition, competing against both specialist AI developers and some of China’s largest technology groups.
Revenue growth at this pace is striking even by the standards of the current AI cycle, however, the impressive figure does not tell the entire story. Zhipu still recorded a net loss of around 2 billion yuan during the first six months of the year, although that was an improvement from approximately 2.4 billion yuan a year earlier.
AI investors are starting to ask a different question
For much of the global AI boom, financial markets have rewarded companies for demonstrating that they have access to the technology, computing infrastructure or intellectual property needed to participate. More recently, however, investors have become increasingly selective about which AI companies can turn spending into sustainable growth. The next stage is likely to be more demanding. Investors increasingly want to see whether enormous expenditure on models, computing power and engineering talent can translate into lasting revenue, and ultimately profits.
Zhipu illustrates this challenge particularly clearly. The company spent approximately 2.1 billion yuan on research and development during the first half, an increase of more than a third from a year earlier. This means that its R&D expenditure alone was more than twice the revenue it generated during the period.
This is not necessarily unusual for a young technology company trying to establish itself in a rapidly developing market. Building competitive AI models requires expensive chips, data centres, specialist engineers and continuous model development. However, the numbers demonstrate why evaluating AI companies purely on revenue growth can be misleading. A business can experience spectacular demand whilst still consuming large amounts of capital.
Zhipu is not the only Chinese AI company growing rapidly
The expansion is also not confined to one company. Chinese AI developer MiniMax recently reported that its first-half revenue had risen by approximately 283%, providing another indication that commercial demand for generative AI products is spreading quickly through China’s technology sector. This is important because investors are beginning to see a broader pattern rather than a single exceptional company.
China has spent several years building an increasingly independent AI ecosystem involving software developers, semiconductor manufacturers, cloud providers and large technology groups. Companies including Alibaba, Huawei and ByteDance are investing heavily alongside younger businesses such as Zhipu, MiniMax, Moonshot AI and DeepSeek.The result is an unusually competitive domestic market.
This level of competition may accelerate innovation and lower the cost of deploying AI, but it also creates pressure on margins. Companies must continually improve their models while competing on price and trying to acquire users before rivals can establish stronger positions. For shareholders, rapidly increasing adoption is therefore only one side of the investment equation.
China is building an AI ecosystem around its own chips
One of the most strategically important developments lies beneath the software itself. Zhipu has been working to reduce its reliance on imported semiconductor technology and recently introduced a lower-cost version of its GLM model that the company says was tested entirely using Chinese-made chips. Such a move carries significance far beyond Zhipu.
Advanced semiconductors have become one of the biggest constraints on China’s AI ambitions. Restrictions on access to some of the world’s most powerful computing chips have encouraged Chinese companies to accelerate development of domestic alternatives and adapt their AI models to work efficiently on locally produced hardware. Progress is increasingly visible across the entire supply chain.
ChangXin Memory Technologies, the Chinese memory-chip producer, has reportedly begun small-scale production of advanced high-bandwidth memory, or HBM, technology. HBM has become particularly important because AI processors require extremely fast access to very large quantities of data. Huawei is also pouring capital into the same technological race.
The company increased first-half research and development expenditure by more than 25% to around 121 billion yuan as it continued investing in AI, semiconductors, smart devices and other advanced technologies. The financial cost is substantial: Huawei’s first-half profit fell sharply even as revenue continued to grow.
Taken together, these developments indicate that China is attempting to build more of the underlying technology stack required to support its own AI industry.
Hong Kong is becoming an important market for the AI trade
For investors, another important change is taking place in Hong Kong. Many of China’s most interesting artificial intelligence companies were previously private businesses, making direct investment difficult for ordinary investors. Public listings are gradually beginning to change that.
Zhipu’s arrival on the Hong Kong Stock Exchange created one of the region’s first opportunities to invest directly in a listed Chinese large-language-model developer, which makes the company particularly interesting from a market perspective.
Until recently, investors who were seeking exposure to Chinese AI generally had to buy shares in diversified technology groups whose businesses extend far beyond artificial intelligence. A pure-play company offers much more direct exposure to AI demand, but also concentrates the risks.
A diversified technology group can use profits from advertising, e-commerce, cloud services or consumer electronics to fund expensive AI development. A specialist AI company eventually has to demonstrate that its core technology can generate enough revenue to support the enormous cost of building it. This is essentially why Zhipu’s results deserve attention beyond the impressive 400% growth figure.
Revenue growth does not automatically justify an AI valuation
AI stocks have repeatedly demonstrated how quickly expectations can become embedded in share prices. When investors believe a technology could transform entire industries, traditional valuation measures can temporarily take a back seat to expectations about future market size. Across the broader AI trade, however, investors have increasingly shifted their attention towards fundamentals, infrastructure and cash-flow generation. But eventually earnings matter.
Evidence increasingly suggests that businesses, governments and consumers will continue adopting the technology, however, the central investment question is how much of that economic value will ultimately flow to the companies building the models.
Competition is already pushing down the cost of using AI. Models are becoming cheaper, open-source alternatives are improving and customers have a growing number of providers from which to choose. This is all positive for adoption but potentially less positive for profit margins.
China may be an especially important test of this dynamic because competition between AI developers is intense and companies are under strong pressure to offer capable models at increasingly low prices. Zhipu therefore represents both sides of the AI investment story: extraordinary growth and extraordinary expenditure.
A broader signal for Asian technology investors
The significance extends beyond individual Chinese AI stocks. AI-related businesses are becoming increasingly important to equity markets across Asia, encompassing semiconductor manufacturing, memory chips, networking equipment, data centres and software. Taiwan, South Korea, China, Japan and parts of Southeast Asia all occupy different positions within the global AI supply chain.
Consequently, the Asian AI investment opportunity is much wider than simply trying to identify the region’s equivalent of the largest US technology companies. Some businesses manufacture the components needed to train AI models and others provide computing infrastructure. Some develop the models themselves, whilst another group is beginning to use artificial intelligence to improve existing industries. The investment characteristics of those companies can be very different.
A profitable semiconductor supplier benefiting from higher AI infrastructure spending should not necessarily be valued in the same way as an early-stage software company that is still spending considerably more than it earns. As the sector matures, investors may increasingly differentiate between them.
Can investors buy Zhipu AI shares?
Zhipu AI is publicly listed on the Main Board of the Hong Kong Stock Exchange under ticker 2513.HK, so investors do not have to be Chinese nationals to own the shares.
In practice, access depends on the investor’s location and brokerage provider. Investors using a broker that offers trading in Hong Kong-listed equities may be able to purchase Zhipu shares in much the same way as other stocks traded on international exchanges. Investors should still check the market access, dealing costs, currency conversion requirements and regulatory restrictions that apply through their own brokerage account and jurisdiction.
The same principle applies to other Chinese technology companies listed in Hong Kong. Access to the shares, however, should not be confused with the suitability of an investment.
Emerging AI companies can experience substantial price volatility, and rapid revenue expansion does not guarantee that a business will become profitable. Valuation, competition, capital requirements and continued technological development all matter.
For investors considering exposure to the sector, understanding those differences may become increasingly important as more Chinese AI businesses enter public markets.
What investors should watch next
Zhipu’s next challenge is relatively straightforward to describe but much harder to achieve: maintain strong revenue growth whilst also bringing expenditure under greater control. Encouragingly, losses narrowed during the first half despite continued investment and growing commercial adoption suggests that Chinese businesses are willing to pay for AI products and services. However, the scale of research spending shows how expensive the race remains.
Investors should therefore watch not only headline revenue growth but also margins, cash consumption, R&D spending and the company’s progress towards profitability. The development of China’s domestic semiconductor industry will also matter because cheaper and more readily available computing infrastructure could eventually reduce one of the largest costs facing local AI developers.
Competition will be another decisive factor. If Chinese AI models continue improving while prices fall, adoption could accelerate substantially. Yet the same competitive pressure could make it harder for individual companies to convert that demand into attractive margins. This may define the next phase of China’s AI investment story.
Zhipu’s 400% revenue increase demonstrates that commercial demand is real. What investors now need to discover is whether the companies at the centre of China’s AI boom can turn that demand into durable earnings.
