Bank of America: China’s cloud market size is expected to reach 1.7 trillion by 2030; reiterates confidence in Tencent and Alibaba

Bank of America Securities’ latest report says that China’s short-term monetization of artificial intelligence (AI) will be driven mainly by enterprise demand. Cloud services are the biggest and clearest commercialized model, covering AI MaaS, AI compute infrastructure, and non-AI cloud. The report expects that, driven by enterprise-level AI application adoption, China’s cloud market will see an all-out explosion, with major internet cloud giants becoming the primary beneficiaries—top picks are Tencent (00700) and Alibaba (09988).

Cloud market size tops at 1.7 trillion by 2030

Bank of America expects that, as AI compute demand surges and MaaS becomes more widespread, the size of China’s cloud market will grow from 389 billion yuan (RMB, same below) in 2025 to 1.73 trillion yuan in 2030, at a 35% compound annual growth rate. Among them, AI cloud business is the core engine, projected to grow rapidly at a 66% compound annual growth rate—rising from 66 billion yuan in 2025 to 1.15 trillion yuan in 2030, accounting for 66% of the overall market.

By breakdown, driven by tokenized (Token) charging models, the AI MaaS market will jump from 21 billion yuan in 2025 to 676 billion yuan in 2030; AI infrastructure cloud will reach 470 billion yuan. Traditional non-AI cloud will move into a mature phase, with a projected size of 583 billion yuan in 2030.

Compute and ecosystem building create moats; cloud giants’ advantages are clear

Bank of America is explicit that it “prefers cloud giants over independent AI Labs.” The report analyzes that, because domestic AI model capability gaps are limited, technical advantages alone are hard to translate into lasting profits. Instead, obtaining compute, ecosystem integration, and distribution channels are the core moats, structurally favoring giants.

Alibaba, Tencent, and ByteDance, leveraging their scale and customer base, will become the main beneficiaries. China’s top four giants’ combined capital expenditures in the first four months of 2025 are close to 400 billion yuan, far exceeding the combined total of Zhipu (02513) and MiniMax (00100). Strong cash flow gives giants an absolute compute advantage, enabling them to tolerate lower model-layer profits to drive cross-selling of cloud services.

Independent labs face compute bottlenecks; accelerating overseas expansion

By contrast, independent labs such as Zhipu and MiniMax, although growing faster, face severe compute bottlenecks—especially for inference—and rely more on external financing. When compute is constrained, they prioritize model training and are highly dependent on giants’ platforms for distribution, facing risks of being bypassed.

To break through, overseas markets have become a key focus for these independent labs to expand. Since early 2026, Zhipu has already had nearly 30% of AI model revenue coming from abroad; Bank of America also expects MiniMax’s overseas API revenue in the first half of this year to contribute more than 30%. Although both are expected to achieve an ARR target of $100 million in 2026, DeepSeek—backed by strong research and development capabilities and not prioritizing profits— is raising industry benchmarks and intensifying competition.

Top picks: Alibaba and Tencent—watch second-half ecosystem-war catalysts

For individual stocks, Bank of America reiterates its “Buy” rating on Alibaba, saying it is the best representative of China’s “AI/Cloud” theme, directly benefiting from the tight compute environment. While recent share prices reflect concerns about weak retail demand, ahead of Alibaba Cloud’s conference in September, its risk-reward ratio is highly attractive, with a target price of HK$168.

Bank of America is also bullish on Tencent, giving it the same “Buy” rating, with a target price of HK$780. With the acceleration of its large models and consumer-grade intelligent agents, cloud revenue is expected to resume growth in the second half of this year. In addition, ByteDance and telecom operators—supported by aggressive capital expenditure—are also variables that cannot be ignored. As major companies roll out model upgrades in the second half (such as ByteDance releasing Seed 2.1 and Tencent releasing Hunyuan 4.0), China’s AI ecosystem war will fully heat up.