Explosive growth! Morgan Stanley: China's AI cloud market will reach a five-year CAGR of 72%, Alibaba is expected to become the biggest winner

```

Morgan Stanley believes that China's AI cloud market (GenAI-related IaaS + MaaS) is experiencing explosive growth.

According to information from Chasing Wind Trading Desk, a Morgan Stanley analyst predicted in a March 16 report that, driven by multiple factors such as surging demand for inference computing power and the accelerated penetration of large model applications, China's AI cloud market will see a compound annual growth rate (CAGR) as high as 72% from 2024 to 2029, with market size jumping from RMB 15 billion in 2024 to RMB 218 billion by the end of 2029.

The proportion of GenAI in China's IaaS + PaaS total market will rise from 6% in 2024 to 39% in 2029, making AI cloud evolve from a marginal player to the core growth engine of the entire cloud computing market.

In this computing power arms race, Morgan Stanley lists Alibaba as the "top pick," giving it an Overweight rating and a US stock price target of $180. Morgan Stanley believes that Alibaba, with its "full-stack" layout from in-house chips (T-Head), GPU infrastructure, Qwen large model, Bailian model platform to 2C/2B application ecosystem, is the best candidate in China's AI infrastructure track, resembling Google's position in the U.S. market. Following closely is ByteDance, whose Volcengine is becoming the most disruptive new entrant in this field. ByteDance's capital expenditure in 2026 is expected to reach RMB 250 billion, surpassing Alibaba's RMB 160 billion and Tencent's RMB 130 billion.

More notably, Morgan Stanley believes that China's cloud computing market is ushering in its first price increase cycle in 20 years. Morgan Stanley estimates that for every 1% increase in AliCloud's prices, its EBITA margin can improve by one percentage point, or push EBITA forecasts up by 11%; if overall contract prices increase by 10% (assuming 20% of contracts are renewed annually), EBITA margin will expand by four percentage points. This profit elasticity is a key variable not yet fully priced in by the market.

Market Size: AI cloud to reshape China's cloud landscape with 72% compound growth

Morgan Stanley cites IDC data indicating that China's overall public cloud market (IaaS + PaaS + SaaS) is about $45 billion in 2024, expanding to $105 billion by 2029, with a CAGR of 18%. IaaS is the largest component, accounting for 56% of the public cloud market; the 2024 IaaS market size is about $25 billion, expected to grow at a 17% CAGR by 2029.

Yet, AI cloud's expansion far outpaces the overall market. IDC forecasts that GenAI's share in IaaS + PaaS will rise from 6% in 2024 to 39% by the end of 2029, becoming the core engine accelerating hyperscale cloud vendors' performance.

From the demand side, inference is the true driver of future growth.

IDC predicts that the share of training workloads in GenAI IaaS market will plunge from 76% in 2024 to 23% in 2029, while inference demand will grow at a CAGR of 103%, far exceeding training's 26%.

Doubao, owned by ByteDance, provides some of the most direct numbers: average daily Token usage soared from 4 trillion in December 2024 to 50 trillion by the end of 2025, more than a 12-fold increase in a year. In addition, MiniMax's annual recurring revenue (ARR) rose from $100 million in December 2025 to $150 million by February 2026, an increase of 50%, with token usage growing sixfold, partly benefiting from explosive adoption of OpenClaw.

It is worth noting that the total size of China's public cloud is only about 10% that of the U.S. (based on 2024 data), with significantly lower penetration rates. Morgan Stanley's CIO survey shows that Chinese companies' willingness to adopt public cloud in the next 12 months has risen from 58% in the previous survey to 71%, indicating huge long-term potential for penetration growth.

Competitive landscape: Alibaba + ByteDance's "dual oligopoly" forms, SOE cloud share under pressure

Morgan Stanley believes that the competitive landscape of China's AI cloud era is undergoing a fundamental restructuring, with two clear trends emerging:

Hyperscale cloud vendors are regaining share from state-owned enterprises (SOEs). 

Between 2021 and 2024, telecom operators and Huawei leveraged aggressive pricing and SOE client support to continuously erode AliCloud's market share. But since the second half of 2024, this trend has reversed—Alibaba Cloud's IaaS market share rebounded from 25.5% in H1 2024 to 26.8% in Q2 2025.

Morgan Stanley attributes this to three main advantages: more innovative AI models, stronger supply chain resource acquisition, and greater execution efficiency in strategic transformation among private enterprises. For instance, starting in 2025, Alibaba, Tencent, and ByteDance's AI-related capital expenditures each exceed RMB 100 billion, while telecom operators during the same period spend only about RMB 20 billion.

ByteDance is the most disruptive new player in the AI cloud era.

Volcengine's share in the GenAI IaaS market has reached 14.2% (Alibaba Cloud leads with 23.5%), and its MaaS market share is as high as 37.5% (1H25), ranking first in the industry. ByteDance's overall share in the public cloud IaaS market has rapidly climbed from near zero pre-2024 to nearly 4% by Q2 2025.

More notably, ByteDance's expected capital expenditure in 2026 is RMB 250 billion, exceeding Alibaba's RMB 160 billion and Tencent's RMB 130 billion. The special attributes of private enterprises mean they tolerate lower short-term margins, enabling aggressive expansion.

In product matrix evaluation, Morgan Stanley measures competitors by supply and capacity, product matrix, and service capabilities. Overall, Alibaba Cloud, as a full-stack provider from in-house chip (T-Head) to foundational model (Qwen) to MaaS platform (Bailian), stands out in comprehensive advantages; ByteDance leads in multimodal models (Seedance 2.0) and the MaaS market; Baidu, despite having a full-stack solution (Kunlunxin chip + Wenxin model + Qianfan platform), lags in cloud market size and model capability, so Morgan Stanley maintains a Neutral rating.

Pricing cycle: First price increase in 20 years, cloud service profit margins at a revaluation threshold

Historically, cloud computing has been a deflationary industry—the larger the scale, the lower the cost, and the more prices fell. But Morgan Stanley points out that the AI era is breaking this paradigm, foreshadowing an unprecedented round of price hikes, with China's AI cloud pricing about to enter its first upward cycle in 20 years.

Global hyperscale cloud vendors are leading the way: On January 4, 2026, AWS raised the price of its machine learning EC2 Capacity Blocks by around 15% (e.g., the p5e.48xlarge instance rose from $34.61 to $39.80 per hour); on January 27, 2026, Google Cloud (GCP) announced steep price increases for network, storage, and AI infrastructure, with some CDN and traffic transmission rates up as much as 100%, starting in May 2026.

Signs of follow-up action have appeared in China: Wangsu Technology announced a 35%-40% CDN price increase starting February 1, 2026; UCloud said it would raise prices for all contract renewals and new clients starting March 1, 2026; Tencent AI platform raised prices for some in-house models by up to 400%. Both companies cited supply chain inflation as the core driver of price increases.

The elasticity of profit margins is noteworthy.

Morgan Stanley estimates that for Alibaba Cloud, every 1% overall price increase can lift EBITA margin by about one percentage point and raise EBITA forecast by about 11% (assuming 20% contract renewals each year, other conditions unchanged); if overall prices rise by 10%, EBITA margin will improve by 4 percentage points.

Morgan Stanley also highlights two unique constraints in China: First, ByteDance still prioritizes grabbing market share, and its aggressive pricing strategy could limit the industry’s overall room for price increases; second, China's lack of independent base model providers with massive computing power demand like OpenAI means supply-side cost pressure is less efficiently passed downstream than in the U.S. market.

Margin improvement: Inference wave + in-house chips + depreciation policy, threefold logic supports marginal improvement

Currently, Alibaba Cloud's EBITA margin is only in the high single digits (F24-25E), far behind AWS (about 35%), Microsoft Cloud (about 40%), and Google Cloud (about 23%). Morgan Stanley believes this gap will gradually narrow as the AI era evolves.

The main drivers are threefold:

The workload shift from training to inference is the most important structural catalyst.

Inference workloads are charged per token/API, can stack higher-value value-added services (VAS), and a single server can process multiple workloads simultaneously (parallel batch scheduling), increasing utilization and improving profit structure. By contrast, training workloads are priced more as commodities, with limited pricing flexibility.

In-house ASICs significantly lower per-unit infrastructure cost. 

Third-party AI chip vendors typically earn gross margins of 50%-60% (chip level) or even 60%-70% (server level), meaning firms with in-house ASICs like Alibaba (T-Head) and Baidu (Kunlun) can procure at over 50% lower cost than competitors, directly reducing inference-side costs.

Depreciation policy adjustments provide additional flexibility for profit improvement.

U.S. hyperscale cloud vendors now depreciate servers over six years, while Alibaba Cloud is still depreciating over 3-5 years; if it follows suit, substantial margin improvement could result.

~~~~~~~~~~~~~~~~~~~~~~~~

The above quality content is from Chasing Wind Trading Desk.

For more detailed analysis, including real-time interpretation and frontline research, please join the [Chasing Wind Trading Desk Annual Membership]

Risk Warning and DisclaimerThe market is risky, and investments must be made cautiously. This article does not constitute personal investment advice, nor does it take into account the special investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, viewpoints, or conclusions in this article suit their particular circumstances. Investment based on this is at your own risk. ```