The global wave of rising AI infrastructure costs has reached China. Following price hikes by Amazon AWS and Google Cloud, Alibaba Cloud and Baidu Smart Cloud both announced on Wednesday (March 18) increases in pricing for AI computing and storage products, with some services seeing hikes of up to 34%. The new rates will take effect on April 18, 2026. Markets responded positively: Alibaba (9988.HK) closed at HK$137.70, up 2.3%, while Baidu (9888.HK) ended at HK$121.80, gaining 2.2%.
According to Alibaba Cloud’s announcement, surging global AI demand and supply chain cost pressures have significantly increased procurement expenses for core hardware. Additionally, its MaaS platform “Bailian” recorded record-breaking token usage growth between January and March, prompting the company to prioritize scarce computing resources for higher-value workloads.
Specific price adjustments include:
- Services based on Pingtouge’s Zhenwu 810E and similar AI accelerator cards: +5% to +34%
- CPFS (AI-optimized file storage): +30%
- Effective date: April 18, 2026
- Grandfathering clause: Existing service contracts remain unaffected during their current billing cycles
Baidu Smart Cloud implemented similar increases on the same day, citing rapidly growing global AI adoption and sharply rising costs for core hardware and related infrastructure:
- AI computing services: +5% to +30%
- Parallel file storage services: +30%
- Effective date: 00:00 Beijing time, April 18, 2026
This round of price hikes reflects a broader industry shift across global cloud providers:
- AWS raised prices by 15% for EC2 machine learning capacity blocks on January 22, breaking its nearly two-decade tradition of only lowering prices.
- Google Cloud announced price increases of up to 100% on data transfer, AI, and compute infrastructure services on January 27.
- Domestic providers like UCloud have also initiated multiple rounds of price adjustments recently.
Tencent President Martin Lau previously noted that beyond GPUs, both storage and CPUs are now experiencing supply constraints. Large cloud vendors, benefiting from scale advantages and long-term supplier contracts, secure priority allocations and then resell capacity to customers.
“For a long time, cloud services in China kept cutting prices—until supply tightened. Now, prices are rising,” Lau remarked. “If supply constraints persist, Tencent Cloud will likely maintain strong pricing power.”
The recent surge in usage of AI agents—particularly the popular “OpenClaw” (nicknamed “Lobster”)—has dramatically accelerated token consumption. Alibaba Cloud’s “Bailian” platform saw daily token usage skyrocket from 30 trillion in mid-2025 to over 180 trillion by Q1 2026.
NVIDIA CEO Jensen Huang recently emphasized that the cost and efficiency of token generation directly determine tech companies’ revenue—and survival. Meta is reportedly considering large-scale layoffs due to mounting pressure from massive AI infrastructure investments.
In a recent research note, Morgan Stanley highlighted that explosive growth in inference-side computing demand and rapid adoption of large models are driving a boom in China’s AI cloud market. The firm named Alibaba its “top pick,” assigning an overweight rating and a US$180 price target for its NYSE-listed shares, citing its comprehensive “full-stack” AI infrastructure—from in-house T-Head chips and GPU infrastructure to the Qwen large model and application ecosystem—as making it the leading play in China’s AI infrastructure race.
