China is preparing to invest approximately 2 trillion yuan ($295 billion) over the next five years to construct a nationwide network of data centers, according to a Bloomberg report published on June 9. Reuters confirmed the report, noting that the initiative will be funded primarily through sovereign debt, specifically ultra-long-term special government bonds of more than ten years’ tenure, and state funds for strategic industry investment, supplemented by bank loans and private capital. The scale of the commitment signals that Beijing is treating AI computing infrastructure as a strategic national priority on par with its earlier investments in high-speed rail and renewable energy.
The plan arrives at a critical juncture. US export controls have progressively restricted Chinese companies’ access to advanced semiconductors from Nvidia and other Western chipmakers, forcing a rethink of how China powers its AI ambitions. Rather than simply seeking workarounds, Beijing appears to be executing a longer-term strategy: build so much domestic computing capacity that the restrictions become less relevant over time.
A Nationwide Buildout Coordinated Across Government and Industry
The infrastructure initiative is designed to be comprehensive rather than concentrated in a handful of tech hubs. While cities such as Beijing, Shanghai, and Shenzhen will remain central nodes, the plan envisions data center clusters spreading into inland provinces where land, water, and renewable energy are more abundant. This geographic diversification serves both economic and security purposes, reducing the vulnerability of China’s AI computing base to any single point of disruption.
The funding model reflects the same public-private coordination that has driven other major Chinese industrial campaigns. Sovereign debt will provide the foundational capital, while state-owned enterprises, including China Mobile and China Telecom, which will operate the bulk of the data centers, are expected to contribute substantial resources. Local governments, eager to attract investment and high-value jobs, will compete to offer favorable conditions for data center construction.
Domestic Hardware at the Center of the Strategy
A defining feature of the plan is its emphasis on equipping new data centers with domestically produced hardware. This is where the infrastructure buildout intersects directly with China’s semiconductor ambitions. Huawei’s Ascend series of AI accelerators is positioned as the primary alternative to Nvidia’s restricted chips, and the new data centers will provide a massive, guaranteed market for these products.
The timing is favorable for Huawei. The company’s latest Ascend 950PR, featured in the Atlas 350 accelerator, delivers 1.56 PFLOPS of FP4 performance and reportedly outperforms Nvidia’s China-specific H20 chip by 2.8 times, according to Zhang Dixuan, head of Huawei’s Ascend computing business. If these performance claims hold up at scale, the case for investing in domestic hardware becomes significantly stronger for Chinese tech companies.
ByteDance has already acted on this logic. The company committed to purchasing over 40 billion yuan ($5.6 billion) worth of Huawei Ascend chips in 2026, with first deliveries slated for early in the year. Alibaba and Tencent have reportedly made comparable commitments. These orders provide Huawei with the revenue and production scale needed to continue improving its chips, creating a virtuous cycle that the new data center plan will accelerate further.
Connecting Infrastructure to the Broader AI Economy
The $295 billion plan is not simply a hardware procurement exercise. It is intended to create the physical backbone for a much broader AI economy. By ensuring that Chinese companies have reliable, affordable, and domestically controlled access to computing power, Beijing aims to remove one of the key bottlenecks that could slow the development and deployment of AI across the economy.
This matters particularly for sectors where China is moving aggressively — including autonomous vehicles, humanoid robotics, industrial automation, and AI-powered healthcare. Beijing has also been pushing data centers to reduce their environmental footprint, with a 29-measure green data center action plan issued in May setting targets for power usage effectiveness and renewable energy sourcing. Each of these industries requires enormous amounts of training compute and inference capacity. A nationwide data center network would provide the infrastructure layer that ties these ambitions together.
The plan also has significant implications for China’s AI model development ecosystem. The country’s leading model developers, including DeepSeek, Moonshot AI, Zhipu AI, and ByteDance’s Doubao team, have all been scaling their compute requirements rapidly. A guaranteed supply of domestic computing resources would reduce their dependence on expensive and increasingly scarce foreign chips, potentially accelerating the pace of model development and deployment.
A Long-Term Bet on Self-Sufficiency
The $295 billion commitment represents a long-term bet that China can build a fully self-sufficient AI technology stack — from chips and data centers at the infrastructure layer, through models and platforms in the middle, to applications and services at the top. Whether that bet pays off depends on whether domestic hardware can continue to close the performance gap with Western alternatives, and whether Chinese software developers can optimize effectively for the new architectures.
The plan also carries a clear message to the rest of the world. By committing to this level of investment, Beijing is signaling that it does not intend to be constrained by export controls indefinitely. It is building the infrastructure to compete regardless of what restrictions Washington imposes — and it is doing so at a speed and scale that few other nations could match.