On March 18, Alibaba Cloud and Baidu Intelligent Cloud revealed price increases for AI computing power, storage, and other offerings, with maximum hikes reaching 34%.

Price Hikes of Up to 34%

At 12:16 PM, Alibaba Cloud's official website posted an announcement stating that due to a global surge in AI demand and escalating supply chain costs, procurement expenses for core industry hardware have risen substantially. Consequently, Alibaba Cloud decided to adjust the pricing of services such as AI computing power and CPFS (Intelligent Computing Version) starting from April 18, 2026.

Specifically, computing card products like the T-Head Zhenwu 810E will see price increases ranging from 5% to 34%, while file storage product CPFS (Intelligent Computing Version) will rise by 30%. Alibaba Cloud clarified that users who have already purchased relevant services before April 18, 2026, will not be affected during the current order cycle. The new prices will take effect starting from the next renewal cycle.

Just a few hours later, Baidu Intelligent Cloud followed suit with its own price adjustment announcement. Due to the rapid expansion of global AI applications, demand for computing power continues to climb, and costs for core hardware and related infrastructure have surged. To ensure long-term stable platform operation and service quality, Baidu Intelligent Cloud will implement structural price optimizations for certain products.

Baidu Intelligent Cloud's adjustments will also come into effect at 00:00 on April 18. AI computing power-related product services will increase by approximately 5% to 30%, while parallel file storage will rise by about 30%.

It is widely believed that Alibaba Cloud is merely the first major provider to raise prices. After Baidu Cloud follows suit, Tencent Cloud may be next, with UCloud and Wangsu Technology potentially following suit. The core reason is that this marks the inaugural year of AI applications in China, with cloud computing and IDC serving as key supporting industries.

According to insiders, a major factor behind Alibaba Cloud's price hike is the "explosive growth in Token calls." During the Spring Festival, AI Agent applications surged, and Alibaba Cloud's MaaS business, BaiLian, achieved record-high growth from January to March this year. Alibaba Cloud is now prioritizing scarce AI computing power resources for Token-related services.

Tokens represent the smallest unit of text processing for AI large models and serve as a key metric for measuring AI application activity. According to the latest data from OpenRouter, the world's largest AI model API aggregation platform, during the week of March 9-15, weekly calls for Chinese AI large models soared to 4.69 trillion Tokens, up 11.83% month-over-month. In contrast, calls for U.S. large models declined by 9.33% to 3.294 trillion Tokens during the same period.

Even more striking is that the top three global models by call volume are all Chinese: MiniMax M2.5 claimed its fifth consecutive championship with 1.75 trillion Tokens, JieYue XingChen Step3.5Flash took second place, and DeepSeek V3.2 secured third with 1.04 trillion Tokens.

JPMorgan predicts that China's AI inference Token consumption will surge from approximately 10 quadrillion in 2025 to about 390 quadrillion in 2030, marking a 370-fold increase over five years.

As Chinese AI models dominate the global stage, demand for computing power has shifted from a steady stream to a tidal wave.

From 'Selling Computing Power' to 'Selling Intelligence'

Behind the price hikes lies a profound transformation in the business logic of Chinese cloud providers. Just two days before the price adjustment announcement, Alibaba completed a major organizational restructuring, officially establishing the Alibaba Token Hub (ATH) business group, directly overseen by CEO Wu Yongming.

This new business group encompasses Tongyi Labs, the MaaS business line, the QianWen business unit, the Wukong business unit, and the AI Innovation business unit, covering the entire spectrum from foundational model R&D and model service platforms to AI applications for individuals and enterprises.

In an internal memo, Wu Yongming stated, "We are on the eve of the AGI explosion. A vast amount of digital work will be supported by tens of billions of AI Agents, which will run on Tokens generated by models and become the primary carrier for human interaction with the digital world." This signifies that Alibaba Cloud, China's largest cloud provider, is shifting its business strategy focus from selling computing power resources to selling intelligence, leveraging its self-developed QianWen large model. Alibaba will build a unified model service platform, BaiLian, through which Alibaba Cloud will provide MaaS commercialization services.

Baidu Intelligent Cloud is also accelerating its deployment. At a recent internal strategy meeting, Baidu Intelligent Cloud executives raised their AI-related revenue growth target for 2026 from 100% to 200%, urging the entire team to strive for high growth and secure the top position in the AI cloud market. Market data supports Baidu's leading position: in 2025, Baidu Intelligent Cloud ranked first in the industry in both the number of winning bids and the total value of AI large model-related tenders, securing the "bidding champion" title for the second consecutive year.

According to media statistics, in 2025, China's mainstream cloud providers won a total of 341 AI large model-related projects, with a total contract value of approximately 2.7 billion yuan. In this fierce competition, Baidu Intelligent Cloud secured 109 projects with a total value of about 900 million yuan.

As early as January 22 this year, Amazon AWS announced a 15% price increase for EC2 instances used in large model training, breaking AWS's nearly two-decade tradition of "only lowering, never raising" prices.

Five days later, Google Cloud announced price adjustments starting May 1 for services including Google Cloud, CDN Interconnect, and AI and computing infrastructure. In North America, data transfer prices increased from $0.04 to $0.08 per GiB, a 100% hike.

From Seattle to Beijing, cloud providers are collectively raising prices, breaking a nearly two-decade-old "industry belief" that cloud services only lower prices, never raise them.

Chai Daixuan, Managing Director of CIC China Insights Consultancy, believes that the recent price adjustments by multiple leading cloud providers indicate that driven by factors such as the surge in AI computing power demand, rising supply chain costs, and deepening service models, price adjustments have become a clear signal for the industry to address challenges in the current development stage.

Does this mean that the cloud computing industry is transitioning from a low-margin era of "price competition" to a value-driven era of "scarcity competition"?