The Chinese cloud computing industry is witnessing a rare reversal in pricing trends. Following Tencent Cloud, both Alibaba Cloud and Baidu AI Cloud have announced price increases, signaling a shift away from the two-decade-long trend of continuous price reductions and heralding the arrival of an unprecedented cycle of rising costs.

On the afternoon of March 18, Alibaba Cloud announced on its official website that, due to a global surge in AI demand coupled with rising supply chain costs, it would adjust prices for products such as AI computing power and storage, with increases of up to 34%. The company's MaaS business, Bailian, achieved its highest-ever growth rate from January to March this year, prompting Alibaba Cloud to allocate its scarce AI computing resources more heavily towards token-based services.

After the A-share market closed on the same day, Baidu AI Cloud also issued a price adjustment notice, increasing prices for its AI computing-related products and services by approximately 5% to 30%, and for services like parallel file storage by about 30%. The fact that two leading providers announced hikes within a span of less than four hours underscores how tight supply and demand for AI computing power is now translating into price pressures.

In a report earlier this week, Morgan Stanley noted that cloud computing has historically been a deflationary industry—where larger scale leads to lower costs and consequently lower prices. The AI era is disrupting this paradigm. With an explosive growth in token usage, an unprecedented cycle of price increases is brewing in China's AI cloud market.

For investors: As inference-side workloads rapidly expand, and constraints in upstream components and data centers push up marginal costs, cloud providers are seeing an increased likelihood of regaining pricing power.

A historically rare event! Leading domestic cloud service providers are raising prices collectively. Alibaba Cloud's price adjustments affect products including AI computing power and storage, with increases of up to 34%. The reasons cited are the global AI demand explosion and rising supply chain costs. Insiders added that the rapid growth in token usage is altering internal resource allocation priorities, leading Alibaba Cloud to direct more of its scarce computing power towards token-related businesses.

Baidu AI Cloud's price adjustments similarly target both AI computing power and storage, with computing products seeing increases of about 5% to 30%, and services like parallel file storage rising by approximately 30%. The fact that similar products saw price hikes from two leading providers on the same day indicates that "tight supply and demand" is moving from affecting delivery times and queue periods to sending a more direct price signal.

Last week, Tencent Cloud had already announced price increases. Its intelligent agent development platform completed optimizations to the billing strategies for certain models. Notably, the input price for the Tencent HY2.0 Instruct model surged from the original 0.0008 yuan per thousand tokens to 0.004505 yuan per thousand tokens, a staggering increase of 463.13%.

Beyond the BAT companies (Baidu, Alibaba, Tencent), other domestic cloud service providers are also taking action. Wangsu Technology announced a 35% to 40% price increase for its CDN products effective February 1, 2026. UCloud announced price hikes for all contract renewals and new customers effective March 1, 2026.

Regarding overseas providers, on January 4, 2026, AWS increased prices for its EC2 machine learning Capacity Blocks by approximately 15%. On January 27, 2026, Google Cloud announced significant price increases for networking, storage, and AI infrastructure, with some CDN and data transfer rates rising by up to 100%. The new prices are set to take effect from May 2026.

Behind the price hikes: An explosion in token consumption. The accelerated penetration of AI Agent applications, exemplified by OpenClaw, has led to a geometric leap in token consumption compared to traditional conversational AI.

A previous UBS report pointed out that the weekly token usage of global AI models, monitored through the API aggregation platform OpenRouter, has recently reached approximately 16 trillion. This is almost triple the level seen in January 2026 (before OpenClaw's launch).

Supply-side constraints stem from chips. Regarding memory chips, the Chairman of SK Group stated at the Nvidia GTC conference that due to systemic bottlenecks in chip production, the global shortage of memory chips is likely to persist until 2030.

For AI chips, rental rates for H100/H200 have increased significantly, with delivery timelines extending into 2027. Simultaneously, newly built computing power centers place higher demands on power supply and liquid cooling systems, resulting in construction and operational costs that are significantly higher than those for traditional data centers.

Amid this combination of exploding demand and constrained supply expansion, cloud infrastructure is more susceptible to entering an inflationary environment characterized by "cost-push pressures and tight supply-demand dynamics."

Is the 20-year "deflationary inertia" loosening? An unprecedented price hike cycle is arriving. Historically, cloud computing has been a deflationary industry—the larger the scale, the lower the cost, and the lower the prices. However, Morgan Stanley notes that the AI era is breaking this paradigm. An unprecedented cycle of price increases is brewing, marking the first upward pricing cycle for China's AI cloud market in 20 years.

Analysts at Morgan Stanley predicted in a March 16 report that China's AI cloud market (GenAI-related IaaS + MaaS) is expected to achieve a compound annual growth rate of 72% from 2024 to 2029. The market size is projected to leap from 15 billion yuan in 2024 to 218 billion yuan in 2029.

The share of GenAI in China's total IaaS + PaaS market is expected to rise from 6% in 2024 to 39% in 2029. AI cloud is transitioning from a "marginal player" to the core growth engine of cloud computing.

Of greater interest is the pricing elasticity on the profit side. Morgan Stanley calculates that for Alibaba Cloud, a 1% price increase could boost EBITA margin by 1 percentage point, or lead to an 11% upward revision in EBITA forecasts. If overall contract prices were to increase by 10% (assuming 20% of contracts are renewed annually), EBITA margin would expand by 4 percentage points.