China's National Development and Reform Commission, along with other institutions, has jointly launched a super infrastructure plan worth 2 trillion yuan. This plan aims to build a nationwide intelligent computing network over the next five years to enhance the international competitiveness of the domestic artificial intelligence industry.
The blueprint sets clear goals for technological independence, requiring the share of local suppliers in key software and hardware to increase to over 80%. This move may lead to a reassessment and downward pressure on the market share of overseas companies like Nvidia (NVDA:US) and AMD (AMD:US) in China.
The project funding will mainly be raised through the issuance of ultra-long-term special government bonds with a maturity of over ten years and strategic industry investment funds. If the integration of the power grid and computing network is successfully achieved, the total expected investment could expand to 5 trillion yuan.
Policy Marginal Changes and Self-Sufficiency Supply Indicators
The nationwide computing hub network blueprint, drafted by the National Development and Reform Commission (NDRC) in conjunction with multiple ministries, is a key component of the top-level design of the six networks proposed earlier this year. To ensure supply chain security and independence, Beijing has set clear quantitative targets, aiming for local suppliers like Huawei, Alibaba (BABA:US), Biren Technology, and Moore Threads to provide at least 80% of intelligent computing chips and related technical support. Nine domestically developed computing chips have already passed official security reviews, laying a compliant foundation for large-scale replacement in key industries and public sectors. If this replacement process proceeds as planned, overseas chip giants may face pressure on their supply chains and marginal revenues in China.
Capital Expenditure Comparison and Low-Cost Factor Advantages
Although the direct investment of 2 trillion yuan over five years appears relatively modest compared to the combined capital expenditure of up to $725 billion by U.S. tech giants like Meta (META:US) and Microsoft (MSFT:US) in the AI field this year, China has significant infrastructure efficiency advantages in labor, component supply chains, and computing center construction costs. Analysts point out that if the supporting power grid upgrades and transmission infrastructure integration projects are included in the combined statistics, the total expected investment scale of this computing infrastructure project could rise to 5 trillion yuan, with its multiplier effect providing long-term support for the domestic capital goods and power equipment sectors.
Government Bond Issuance and Diversified Fund Allocation
In terms of funding support and fiscal tool selection, the central government will primarily rely on issuing ultra-long-term special government bonds to provide underlying cash flow. These debt instruments generally have a maturity of over ten years, aiming to match the construction cycle of long-term digital assets. Additionally, government-led strategic industry investment funds will act as the main co-investors, in conjunction with special low-interest loans from state-owned commercial banks, leveraging this to attract private capital participation. This diversified financing combination is expected to provide stable funding chain support for data center connectivity and subsequent operations without crowding out market liquidity.
Computing Resource Integration and Cross-Regional Industry Transmission
According to the planned schedule, the Chinese government aims to integrate the currently scattered and fragmented regional data facilities into a cohesive and efficient network by 2028. State-owned telecom operators like China Mobile (600941:CH) and China Telecom (601728:CH) will primarily be responsible for the daily operation and data transmission assurance of these new computing hubs. The establishment of a unified computing network will effectively reduce the digital transformation costs for high-energy-consuming and high-computing-demand enterprises in finance, intelligent manufacturing, and modern logistics. If resource allocation efficiency is released as planned, the marginal reduction in computing costs is expected to accelerate the iteration speed of domestic large models and guide digital industry investment and technical talent to achieve a gradient shift to energy-rich inland provinces.
