MaaS 到底给中国云厂商带来了什么？｜产业深度

This article outlines the latest developments in China's cloud computing industry in 2026, focusing on how AI and Model-as-a-Service (MaaS) are reshaping cloud vendors' business. Key takeaways are as follows:

1. The industry currently shows a pattern of localized AI growth amid overall stable market expansion: Leading cloud vendors are seeing explosive growth in AI-related business. For example, ByteDance's Volcano Engine reports a 16x year-over-year increase in large model call volume, and AI revenue now accounts for over 30% of Alibaba Cloud's total revenue for the first time. However, the overall growth of China's public cloud market remains only 8%-11%, far lower than the double-digit growth rates seen during the industry's golden age.

2. The explosive growth of AI cloud comes primarily from three sources: bulk GPU procurement by large model startups, premium pricing for high-end GPU tuning services, and internal settlement demand driven by business upgrades at leading internet companies.

3. MaaS has emerged as a new growth driver for cloud vendors. It delivers new revenue through deep customer lock-in, low-threshold customer acquisition and cross-selling of complementary products, and has already become the new benchmark for value measurement in the cloud computing industry.

This article analyzes the current development status of the AI cloud and MaaS industry, offering actionable insights for brands planning AI transformation and navigating consumer trends. Key takeaways are as follows:

1. Traditional enterprise clients remain highly cautious about adopting AI cloud, primarily due to unresolved large model hallucination issues and unproven return on investment (ROI) for AI projects. Brands prioritizing AI transformation should first address concerns around data security, system stability and ROI validation, and avoid large-scale blind investment.

2. Current AI growth is concentrated in pioneering sectors including large model startups, the internet industry and autonomous driving, while demand from traditional industries has not yet been fully unlocked. Brands can adopt a "test and learn" strategy to advance AI transformation incrementally.

3. The MaaS model significantly lowers the barrier for brands to adopt AI. Instead of investing heavily in building in-house computing power and models, brands can test AI applications in marketing, production, customer service and other scenarios through a lightweight pay-per-token model, and purchase complementary cloud products such as storage and security services from cloud vendors to keep transformation costs low.

This article analyzes the current development status, opportunities and risks of the AI cloud track, offering high-value reference for sellers expanding into AI-related businesses. Key takeaways are as follows:

1. Clear growth opportunities exist in the AI and MaaS track: IDC forecasts that total token consumption in China's MaaS market will reach approximately 40 quadrillion by 2026, marking extremely rapid market growth. Leading vendors have already raised their MaaS revenue targets, and the sector has shifted from strategic investment to scalable revenue generation.

2. Small and medium-sized sellers and independent developers have gained new access to the market: MaaS converts capital-intensive GPU computing power into lightweight API access, with the pay-per-token model lowering entry costs to less than 0.1 Chinese cent per call. This has enabled long-tail customers that could not previously afford tens of thousands of yuan in monthly server costs to launch AI-related businesses.

3. Market participants should note key risks: AI cloud is a capital-intensive industry with long payback periods. Demand from traditional enterprise clients has not yet scaled, and current growth is concentrated in pioneering sectors. Sellers should control investment reasonably and avoid blind expansion.

This article analyzes the AI-driven transformation trend in the cloud computing industry, offering key insights for traditional factories advancing digital and AI transformation. Key takeaways are as follows:

1. Traditional factories can adjust their AI transformation strategies based on current industry characteristics: Traditional enterprises remain generally cautious about adopting AI cloud, mainly due to unresolved large model hallucination issues and unproven ROI for AI projects. Factories can prioritize testing AI in small-scale use cases, scale investment only after validating ROI, and avoid risks from excessive one-time upfront investment.

2. The MaaS model lowers the barrier to AI transformation for factories: Instead of bearing high upfront costs to build in-house AI data centers and develop custom large models, factories can test AI applications in product design, production scheduling, supply chain management, customer service and other scenarios through lightweight pay-per-token access, drastically cutting the capital required for transformation.

3. Leading cloud vendors' MaaS platforms are now fully mature, offering complementary services such as vector databases and security protection. Factories can leverage the infrastructure of these mature platforms to advance their digital AI upgrade, without building the entire system from scratch.

This article provides an in-depth analysis of development trends, customer pain points and market opportunities in the AI cloud and MaaS industry, offering high-value reference for technology service providers planning their business布局. Key takeaways are as follows:

1. The overall industry development trajectory is clear: The cloud computing industry has entered a new AI-driven stage, where the industry narrative and value metrics have shifted from traditional CPU core counts to GPU utilization and token call volume. MaaS has become the core growth engine for cloud vendors, with enormous overall market expansion potential. By 2026, total token consumption is projected to reach 40 quadrillion, putting the industry in a period of rapid growth.

2. Core customer pain points in the current market are clear: The core demands of traditional enterprise clients adopting AI are system stability and data security. With unresolved large model hallucinations and unclear ROI for AI projects, most traditional clients hesitate to make large-scale investments. Service providers can develop targeted solutions addressing these pain points.

3. Clear market opportunities exist: The MaaS model has driven rising demand for complementary products including vector databases, secure cloud products and high-throughput object storage. Service providers can develop vertical supporting services around the MaaS ecosystem to access new high-growth markets.

This article analyzes the latest developments in China's cloud platform industry and summarizes the operational advantages of the MaaS model, offering insights for business development for all types of platforms. Key takeaways are as follows:

1. Market demand has shifted: Enterprise demand for cloud services is gradually shifting from traditional general-purpose CPU computing to AI computing power and MaaS services. Leading platforms have already adjusted their business focus and disclosure priorities, so platforms need to keep up with industry changes and布局 AI-related businesses in advance.

2. The MaaS model helps platforms improve operational efficiency and customer retention: After enterprises access a MaaS platform to develop custom models, deep integration of their data and model parameters makes migration to other platforms extremely difficult, allowing platforms to secure long-term stable recurring revenue.

3. Actionable operational takeaways: Platforms can use the low-threshold pay-per-token pricing model to attract long-tail customers including small and medium-sized enterprises and individual developers, while driving cross-selling of high-margin core cloud products through MaaS-led customer acquisition to boost overall revenue.

4. Key risks to note: AI businesses require large upfront capital investment with long payback periods. Platforms should plan capital expenditure reasonably and avoid over-investing ahead of demand.

This article provides a comprehensive overview of the latest industry developments in China's cloud computing sector in 2026, and analyzes the business model transformation brought by MaaS, offering high value for industry research. Key findings are as follows:

1. The industry has taken on new structural characteristics: The core industry narrative of China's cloud computing sector has fundamentally changed in 2026, with leading cloud vendors shifting their disclosure focus to AI-related metrics. The industry now has a structural pattern of "explosive localized AI growth amid slowing overall market expansion": from 2025 to the first half of 2026, overall public cloud growth stood at only 8%-11%, far lower than historical growth rates.

2. The sources of AI cloud growth have a new structure: The current explosive growth of AI cloud comes primarily from bulk GPU procurement by large model startups, premium pricing for high-end GPU tuning, and internal settlement from business upgrades at leading internet giants. AI demand from traditional government and enterprise clients has not yet been unlocked at scale, so growth features high concentration and high average contract value.

3. The industry has seen innovative business model development: MaaS breaks the old cycle of selling raw computing power and price competition that defined traditional cloud computing. It creates a new growth path through deep customer lock-in, low-threshold customer acquisition and cross-selling, reshapes the value measurement standard for the cloud computing industry, and forms a new industrial growth paradigm.