Nomura: China's RMB 2 Trillion AI Infrastructure Investment Benefits Local Supply Chain

Overview

This flash note comments on media reports regarding China's plan to invest approximately $295 billion (about RMB 2 trillion) in building a national data center network over the next five years. Nomura points out that although this investment framework is not a new concept, the news may still serve as a short-term catalyst for the domestic AI supply chain given the government's firm determination to build an independent and controllable AI value chain. The report outlines three beneficiary directions—computing power, networking, and IDC—while also highlighting key risks such as chip capacity bottlenecks and monetization pressure at the application end.

Core views

Policy and Event Level: According to Bloomberg, the National Development and Reform Commission and other departments are drafting a blueprint to build an interconnected computing hub network nationwide. State-owned enterprises such as China Mobile and China Telecom will be responsible for operating most data centers and ensuring network connectivity. In terms of key technologies, the government's strategy clearly relies on local suppliers, including Huawei, to provide core components such as AI chips. If supporting infrastructure investments such as power networks are included, the total investment scale could reach RMB 5 trillion. Nomura emphasizes that although similar discussions occurred as early as January 2025, the disclosure of specific amounts this time reinforces the signal significance of policy implementation. Beneficiary Targets and Industry Chain Logic: If the government and private sector increase infrastructure investment in the coming years, the report believes the main beneficiaries will be concentrated in three areas. First is the AI computing/server sector, such as Unisplendour Corporation; second is the AI networking sector, covering optical transceivers, components, network switches, and optical fibers, represented by companies like Yangtze Optical Fibre and Cable; third is the IDC/cloud service sector, including China Mobile, China Telecom, GDS Holdings, and VNET Group. These companies directly correspond to the incremental demand for computing power infrastructure construction. Risk Warning: Although infrastructure investment brings short-term catalysis, Chinese AI supply chain stocks face two core constraints. First is the lack of advanced AI chip capacity, which may limit the actual speed of computing power expansion; second is the increasing competitive pressure from enterprise and consumer application markets, which may hinder the commercialization and monetization strategies of related companies, leading to the risk of "infrastructure without profitability."

Analysis framework

The report adopts a typical "event-driven + industry chain mapping" analysis approach. It first confirms key factual elements of the news event, such as investment scale, executing entities, and technical routes; then, combined with the government's long-term strategic intent for AI independence, it judges the short-term catalytic effect of the event on market sentiment; finally, it breaks down the macro investment amount into specific beneficiary sub-industries and listed companies along the physical chain of AI infrastructure construction (computing hardware → network connection → data center operations), while assessing potential risks from the dimensions of supply-side bottlenecks and demand-side monetization.

Methodology notes

• Industry/Sector Analysis FrameworkUpstream-Midstream-Downstream Transmission

  Identifying beneficiary links along the AI infrastructure construction chain

  The report breaks down the RMB 2 trillion macro investment plan into three specific beneficiary sectors: AI servers (upstream computing power), optical modules/network equipment (midstream transmission), and IDC/cloud (downstream operations). This helps investors understand how huge capital expenditures translate into orders and revenue opportunities for companies in different links along the industry chain.

• Event Gaming and Behavioral FinanceEvent-driven analysis

  Distinguishing the marginal impact of new information versus old expectations

  The report clearly states that the RMB 2 trillion investment plan is not entirely new news (discussions existed in January 2025), yet still defines it as a "short-term catalyst." This reflects the core logic of event-driven analysis: even if the information itself is not new, when details such as specific amounts and executing entities are finalized, it can still strengthen market confidence and trigger trading opportunities. The key lies in the marginal change in expectation gaps rather than the absolute novelty of the information.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• Unisplendour Corporation (000938.SZ)

  Core beneficiary in the AI computing/server sector

  Risks

  Insufficient capacity of advanced AI chips may constrain server shipments

• Yangtze Optical Fibre and Cable (6869.HK)

  Beneficiary in the AI networking sector, covering optical fibers/modules/switches

• China Mobile (941.HK) / China Telecom (728.HK)

  State-owned enterprises leading data center operations and network connectivity

  Strengths

  Undertake the main operational responsibilities for the national computing hub network

• GDS Holdings (GDS.US) / VNET Group (VNET.US)

  Third-party IDC/cloud service providers benefiting from computing infrastructure expansion

  Risks

  Intensified competition in enterprise and consumer application markets may hinder monetization

Key data

• Direct Investment Scale in AI Infrastructure$295 billion (approx. RMB 2 trillion)For national data center construction over the next five years
• Total Investment Scale Including Supporting FacilitiesRMB 5 trillionEstimated total if infrastructure investments such as power networks are included
• Investment Period5 yearsTime span of the national data center network construction plan

Impact & implications

For China's AI industry, this investment plan signifies a further elevation in the national level of importance attached to computing power infrastructure, helping to accelerate the verification and iteration of key technologies such as domestic AI chips. For related listed companies, there is hope for clear order expectations and valuation repair momentum in the short term, especially for enterprises with state-owned backgrounds or those deeply bound to local leaders like Huawei. However, in the medium to long term, whether infrastructure investment can be converted into sustainable commercial returns still depends on breakthroughs in chip supply bottlenecks and the maturity of the application ecosystem.

Risks

• Lack of advanced AI chip capacity may limit the actual construction progress of computing power infrastructure
• Intensified competition in enterprise and consumer application markets may hinder the commercial monetization capability of the AI industry chain