Citigroup: Market concerns over China's plan to invest 2 trillion yuan in data centers may be premature.

date
10:05 11/06/2026
avatar
GMT Eight
The industry believes that government intervention is more likely to promote wider AI applications, as many public and small and medium-sized enterprises may not be able to afford the high costs of token usage. Therefore, government-funded capabilities will mainly serve state-owned enterprises and smaller businesses.
Citibank's research report stated that after institutions reported that China plans to invest approximately 2 trillion yuan in data centers over the next five years to promote the domestic AI industry, the stock price of BABA-W (09988) in the Hong Kong market has fallen. The market may be concerned that the government's planned data centers will affect the pricing and future investment returns of private super-large-scale enterprises. The bank believes that these market concerns may be premature, and the government's investment should not be seen as a zero-sum game with private super-large-scale enterprises such as Alibaba, Tencent (00700), and BIDU-SW (09888). The bank believes that government intervention is more likely to promote a wider application of AI, as many public and small enterprises may not be able to afford high token usage costs, so government-supported capacity will mainly serve state-owned and smaller enterprises. Citibank also pointed out that this will enable Alibaba and other private participants to focus on serving high-margin enterprise customers that require exclusive support and frequent updates of AI solutions. In addition, super-large-scale enterprises like Alibaba can also lease this capacity to alleviate initial capital spending pressures and reallocate funds to support the innovation research and development of chips, models, and solutions. Due to government investments in building high-speed computing power channels and backbone network interconnection, enterprises like Alibaba may also benefit from lower transmission latency and bandwidth costs between their data centers (IDC) and business nodes.