这项投资将通过主权债务进行融资，包括超长期特别国债（通常期限超过 10 年）、国家战略投资基金、银行贷款和民间资本。将电网与该项目整合可能会使总投资额提升至至少 5 兆元人民币（约合 7380 亿美元）。虽然 2 兆元人民币的数字非常庞大，但它仍小于像 Meta Platforms Inc. 和 Microsoft Corp. 这样的美国科技巨头仅在 2026 年就计划在人工智慧（AI）上投入的超过 7000 亿美元。然而，由于较低的劳动力、组件和建设成本，以及地方政府的激励措施，中国的数据中心建设成本低于美国。
根据 Forrester Research 的首席分析师 Charlie Dai 的说法，统一的计算网络将汇集零散的区域计算资源，并加速人工智慧（AI）模型的迭代以及代理型和物理型人工智慧（AI）服务。尽管 Bloomberg Intelligence 的分析师 Robert Lea 指出，商业化变现仍面临挑战，且主要收益将流向更广泛的经济体而非民营开发商，但像华为这样的本土供应商将受益最多。此外，华盛顿虽然批准 Nvidia 向中国客户销售其前一代的 H200 人工智慧（AI）晶片，但由于北京对本土替代方案的信心不断增强，这些组件尚未开始出货。值得注意的是，来自华为、阿里巴巴、壁仞科技和摩尔线程等公司的九款本土人工智慧（AI）晶片最近通过了国家安全审查。
China plans to spend approximately 2 trillion yuan (approx. $295 billion USD) over the next five years to build a unified nationwide network of data centers. Key government agencies like the National Development and Reform Commission (NDRC) are drafting the blueprint. State-owned enterprises like China Mobile Ltd. and China Telecom Corp. will operate the majority of these centers. To decrease reliance on foreign technology, the plan aims to source at least 80% of AI chips and technologies from domestic suppliers like Huawei Technologies Co., effectively phasing out American companies like Nvidia Corp. and Advanced Micro Devices Inc. (AMD).
This investment will be funded through sovereign debt, including ultra-long-term special government bonds (typically with tenures exceeding 10 years), state strategic investment funds, bank loans, and private capital. Integrating the power grid with this project could elevate the total investment to at least 5 trillion yuan (approx. $738 billion USD). While the 2 trillion yuan figure is substantial, it is smaller than the over $700 billion USD that U.S. tech giants like Meta Platforms Inc. and Microsoft Corp. plan to spend on AI in 2026 alone. However, lower labor, component, and construction costs, along with local incentives, make data center construction in China less expensive than in the U.S.
According to Charlie Dai, principal analyst at Forrester Research, a unified computing network will pool fragmented regional computing resources and accelerate AI model iteration and agentic/physical AI services. Although Robert Lea of Bloomberg Intelligence notes that monetization remains challenging and primary benefits will flow to the broader economy rather than private developers, local suppliers like Huawei stand to benefit most. Additionally, Washington's approval for Nvidia to sell its previous-generation H200 chips has not yet translated into shipments, as Beijing gains confidence in domestic alternatives. Notably, nine homegrown AI chips from companies including Huawei, Alibaba, Biren Technology, and Moore Threads recently passed national security reviews.