DataCanvas Plans Hong Kong IPO: 9-Fold Revenue Growth in Two Years, Third-Party Intelligent Computing Faces an Uphill Battle to Stand Out

Shift from Basic Software to Computing Power Services

On September 29, Beijing Alaya Cloud Technology Co., Ltd. (abbreviation: Alaya Cloud) submitted a listing application to the Hong Kong Stock Exchange, with China Galaxy International Securities (Hong Kong) Co., Ltd. as the sole sponsor, targeting the title of "the first stock of new intelligent computing cloud".

Expanding from AI basic software to intelligent computing cloud, the company has brought a business with heavier investment and more attention to its growth to the capital market.

The Transformation of Software Company

Founded in 2013, Alaya Cloud started its business with data science platforms and AI basic software in the early stage of development, helping enterprises develop, manage and deploy models.

After the rise of large models, customers no longer only need development tools, but also require training computing power, inference support and cross-device scheduling. Following this path, Alaya Cloud has extended its business to intelligent computing operating systems, intelligent computing cloud and computing power services.

At present, Alaya Cloud has products including Alaya NeW Cloud, Alaya NeW OS, and "computing power package". Among them, the intelligent computing cloud provides model training, inference and management services for enterprises and developers; the operating system undertakes underlying work such as computing power scheduling.

This product structure has been reflected in its revenue. From 2023 to 2025, the operating revenue of Alaya Cloud increased from 106 million yuan to 1.098 billion yuan, with a CAGR of 935.85%; in the first half of 2026, it rose 92.0% year on year to 588 million yuan.

However, the surge in revenue does not mean that the business model has been fully validated. This listing application cannot be simply interpreted according to the story of an ordinary software company. The value of software products is usually reflected in whether customers are willing to pay continuously for functions and services; while intelligent computing cloud also has to face equipment investment, idle resources, power and operation and maintenance costs.

The more customers use the services, the faster the revenue may grow; but if the procurement of computing power and project expansion outpace the demand, the capital pressure will also increase accordingly.

In 2025, the adjusted net loss of Alaya Cloud was 91.27 million yuan, which narrowed to 14.88 million yuan in the first half of 2026. However, the narrowing of loss came from the surge in revenue, and continuous investment is still required for computing power procurement and project expansion.

Since its establishment, Alaya Cloud has obtained multiple rounds of financing, the last round of which was completed in June 2026 with a total fundraising amount of 500 million yuan.

Capital has supported the expansion of technology and business, while listing means another round of test: how much stable revenue the past investment has generated, and how much more capital will be needed for future growth.

How Computing Power Transactions Are Realized

Alaya Cloud proposed the measurement method of "Degree Computing Unit" (DCU) and launched the computing power package, hoping to turn the complex computing power procurement into an easier-to-understand, on-demand service.

For an enterprise that is just starting to develop AI applications, this idea has practical appeal. It may only need to train the model once in a short term, but run inference tasks for a long time; if it purchases equipment for peak demand on its own, a large amount of resources may be idle at ordinary times.

Alaya Cloud tries to gather tasks from different customers through the cloud platform, improve equipment utilization efficiency through scheduling, and deliver computing power together with development tools.

What customers value is whether they can start work as soon as possible, whether the task operation is stable, and whether the final cost is cost-effective. The measurement method can lower the purchase threshold, but whether customers will continue to pay depends on the actual experience and price.

Local intelligent computing centers are another scenario for Alaya Cloud to demonstrate its delivery capabilities. The Huangshan "Dawei" Intelligent Computing Center, which Alaya Cloud cooperated with Huangshan Tourism Group, has an estimated project investment of 260 million yuan, and large model services have been launched online.

It should be clarified that the 260 million yuan is the estimated project investment, which cannot be regarded as the revenue of Alaya Cloud or its own capital expenditure.

How much computing power is paid for use after the project is completed, and how the partners share the costs, will determine how much operating value it can bring to the company.

Regional market data also provides an observation perspective. According to the report released by Analysys, based on the caliber of the inclusive intelligent computing cloud market for third-party providers in East China in the second half of 2025, Alaya Cloud's market share is 10.2%.

This data shows that Alaya Cloud has gained a presence in specific market segments, but it cannot be extrapolated to the national intelligent computing market share, let alone directly used to infer its profitability.

Data from CIC shows that in terms of 2025 revenue, Alaya Cloud ranks 9th among intelligent computing service providers in China, and is the only new intelligent computing cloud provider among the top 10 players.

Market share is the result of competition at a specific point in time, while customer renewal and cash collection are the tests over a longer period of time.

From products to projects, Alaya Cloud has built a relatively complete business framework. Next, the capital market will break down this framework for review: what is the proportion of cloud service revenue and project revenue respectively? Can the added value brought by software improve the overall gross margin? After a regional project is verified, how much manpower and capital does the company need to invest to replicate it in other regions? These questions are closer to the essence of the business than "how much computing power the company owns".

Growth Needs Rational Calculation

The opportunities for intelligent computing cloud are extending from model training to inference. Enterprises need to use computing power intensively when developing models, and after the models are embedded into products, more frequent invocation demands will be generated.

Alaya Cloud is engaged in both underlying scheduling and cloud services. If it can enable different tasks to share resources and reduce idle operation, it will have the opportunity to turn its technical capabilities into cost advantages.

However, the challenges are also very specific.

Large cloud vendors have advantages in capital, infrastructure and ready-made customers, and other intelligent computing service providers are also competing for price-sensitive development teams. Customers can compare the performance and quotation of different platforms, while suppliers have to take responsibility for equipment update, stable operation and continuous service.

When the price of computing power drops, the growth of usage volume may not be synchronously converted into profits; the emergence of new equipment will also change the economic value of existing equipment.

In addition, local projects have their own operating rhythm. Signing, construction, delivery and actual use by customers usually do not happen at the same time.

The successful launch of a center proves that the company has the implementation capability; whether the center can obtain enough paid tasks after completion determines whether the investment can be recovered continuously.

For intelligent computing cloud enterprises, the number of projects, the scale of managed computing power and the quality of revenue must be considered separately.

Therefore, the most noteworthy part of Alaya Cloud's listing application is not the grand computing power target, but the continuous operating data: what businesses contribute to the revenue growth, how the gross profit margin changes, whether accounts receivable and operating cash flow can keep up with the revenue, whether the main customers are highly concentrated, and whether the utilization of computing power resources can support the next round of expansion.

Only by combining these figures can we answer whether Alaya Cloud has moved out of the stage of relying on financing to drive scale growth.

Alaya Cloud's transformation has captured a real demand in the artificial intelligence industry: enterprises need to access and use computing power more conveniently.

After submitting the listing application, investors will not only focus on how much computing power Alaya Cloud can schedule, but also see how long customers are willing to use its services, how much profit can be retained from each revenue, and what cost the company has to pay for growth.

This article is from the WeChat Official Account "Insight IPO" (ID: dongcha-ipo), written by Yang Licheng, edited by Chen Xiaoran, and authorized for release by 36Kr.