As the National Unified Computing Power Network becomes increasingly dense and fully developed, computing power will truly become as accessible and efficiently circulated as water and electricity, providing a continuous and powerful source of momentum for the construction of Digital China. 随着全国一体化算力网不断织密完善，算力将真正实现像水、电一样便捷取用、高效流通，为数字中国建设注入源源不断的澎湃动力。
Liu Tong (刘彤), “Strengthen Computing Power New-Type Infrastructure to Consolidate the Foundation of Digital China” (夯实算力新基建 筑牢数字中国底座), People’s Posts and Telecommunications (人民邮电), March 13, 2026, Page 2
A recent Chinese policy discussion on computing power reveals how Beijing now understands the future of artificial intelligence. AI is no longer framed as a standalone technological race, but as a demand signal for a new kind of national infrastructure. What is emerging is a coordinated system, one that integrates computing resources, energy systems, and governance mechanisms into a unified architecture. This is the compute layer of Digital China.
China’s approach is undergoing a quiet but consequential shift. The decisive factor is no longer simply progress in algorithms or applications. Instead, Chinese policymakers are increasingly defining success in terms of the systems that make those technologies possible, above all, the ability to build, coordinate, and govern computing power at national scale. This system is increasingly formalized in what Chinese policymakers call the National Unified Computing Power Network.
This shift reflects a deeper change in how Beijing understands competition in the digital age. The decisive factor is not simply who leads in individual technologies, but who can construct the underlying infrastructure that allows those technologies to operate, scale, and deliver value across an entire economy. In this context, computing power is not just a technical input. It is becoming a strategic resource.
In effect, China is building a national compute system: one that integrates infrastructure, energy, scheduling, and markets into a single coordinated architecture.
This analysis is based on Liu Tong’s recent essay in People’s Posts and Telecommunications News: “Strengthen Computing Power New-Type Infrastructure to Consolidate the Foundation of Digital China.” The essay examines China’s evolving approach to computing power and its role in the construction of Digital China following new central guidance released in the 2026 Government Work Report and 15th Five-Year Plan.
New Infrastructure to National System
China has spent the past several years building out its computing power capacity at scale. By the end of 2025, the country’s computing power market exceeded 835 billion yuan (121 billion USD), with intelligent computing capacity surpassing 1,590 EFLOPS. These figures place China second globally in both total and AI-focused computing power.
Yet the significance of this development lies less in the numbers than in what they enable. Policymakers increasingly describe the current phase as a transition from “scale expansion” to “value release.” The goal is no longer simply to build more infrastructure, but to ensure that computing power can be effectively utilized: delivered where it is needed, when it is needed, and at a cost that supports broad adoption.
This shift marks the emergence of a more structured system. General-purpose computing, intelligent computing, and supercomputing are no longer treated as separate domains, but as components of a coordinated architecture. The emphasis is on integration, not accumulation. What is taking shape is not just capacity, but a functioning national system.
Emergence of a National Compute Network
At the center of this transformation is the development of a National Unified Computing Power Network (also sometimes abbreviated National Compute Network). This network is not simply infrastructure. It is a system for organizing and allocating computing power at national scale. The objective is straightforward but ambitious: to connect computing resources across regions, platforms, and providers into a unified system that can be monitored, scheduled, and accessed on demand.
This effort reflects a recognition that scale alone is insufficient. Without coordination, large amounts of computing capacity can remain underutilized or inefficiently distributed. Policymakers have described the resulting problem in simple terms: “computing power can be hard to find, difficult to dispatch, and poorly used.” The solution is to build a system that makes those resources visible, accessible, and allocable at national scale.
The proposed architecture, described as a “1+M+N” system, suggests a layered design, combining national coordination with regional hubs and distributed resource nodes. Over time, this structure is intended to transform fragmented infrastructure into a coherent platform. In this model, computing power begins to resemble a networked service, rather than a fixed asset tied to a particular location.
Compute–Energy Nexus
As this system expands, a fundamental constraint has come into focus: energy. The relationship is captured in a widely cited phrase: “the end of AI is computing power, and the end of computing power is electricity.” The implication is clear. The limits of artificial intelligence are increasingly defined not by software, but by the availability and cost of power.
Computing power is inherently energy-intensive. Large-scale data centers consume electricity at levels comparable to entire urban populations. This creates a structural imbalance within China’s system. Demand for computing power is concentrated in the eastern regions, while energy resources are more abundant elsewhere. The result is a mismatch between where computing is needed and where it can be most efficiently powered.
In response, policymakers are advancing a strategy of Computing Power-Electrical Power (or Electricity) Coordination (also abbreviated Compute–Energy Coordination). This includes both relocating energy-intensive workloads to resource-rich regions and expanding power generation, transmission, and storage to meet demand where it exists. New approaches such as direct connections between computing centers and energy producers suggest a deeper integration of digital and physical infrastructure. The development of computing power is increasingly inseparable from the development of the energy system that sustains it.
Fragmentation to Orchestration
Even as capacity grows, a new challenge has emerged: fragmentation. The rapid expansion of computing infrastructure has produced a distributed ecosystem spanning regions, providers, and technical architectures. Without coordination, this diversity becomes inefficiency.
Policymakers have been unusually direct in describing the problem. Computing resources, they note, are often difficult to locate, allocate, and use effectively. These are not problems of scarcity, but of organization. The system has reached a scale where coordination is now the central issue.
The response is a shift toward orchestration. Efforts to build interconnection nodes, improve interoperability, and strengthen national-level scheduling aim to ensure that computing power can move across the system as needed. The objective is to create a governance layer above the infrastructure itself, one that enables dynamic allocation and efficient use. This marks a transition from building assets to managing flows.
Emergence of a Compute Economy
As the system matures, a new concept is taking shape: the computing power economy (also abbreviated compute economy). Unlike earlier phases of the digital economy, which were characterized by declining marginal costs, computing power is constrained by physical inputs: hardware, energy, and infrastructure. These constraints introduce a different economic logic.
In this emerging system, computing power must be priced, allocated, and traded. Policymakers have begun to highlight the absence of fully developed market mechanisms, including price formation, transaction matching, and pricing indices. Addressing these gaps is increasingly seen as essential to improving efficiency and supporting broader adoption.
Proposals now include standardized pricing based on units such as “yuan per million tokens,” as well as the creation of a National Computing Power Exchange. These measures point toward the development of a formal market infrastructure for computing resources. In this model, computing power is not simply consumed, it is transacted, governed, and integrated into the broader economic system.
Policy Alignment and System Consolidation
One of the most important signals in this development is the consistency of policy language across major Party-state documents. Key concepts, such as Ultra-Large-Scale Intelligent Computing Clusters, Compute–Energy Coordination, and National-Level Monitoring and Scheduling appear explicitly in both the 2026 Government Work Report and the 15th Five-Year Plan.
This alignment is significant. In China’s policy system, Five-Year Plans define long-term structure, while Government Work Reports drive near-term implementation. When the same technical language appears across both, it typically indicates that a priority has moved beyond experimentation and into coordinated national execution.
The result is a high degree of system coherence. Computing power is no longer treated as a sectoral issue, but as a foundational element of national strategy. It is being embedded simultaneously in planning frameworks, policy instruments, and implementation mechanisms.
End-State Vision: Compute as a Utility
The direction of travel is increasingly clear. Chinese policymakers envision a system in which computing power is as accessible and widely available as traditional utilities. The analogy to water and electricity is explicit, and it is instructive.
In this model, computing power is not scarce or specialized. It is ubiquitous, standardized, and delivered on demand. It flows across regions and sectors, supporting economic activity in the same way that energy and infrastructure have in earlier stages of development.
Achieving this vision requires the integration of multiple systems, computing infrastructure, energy networks, and market mechanisms, into a unified whole. It also requires governance structures capable of coordinating these elements at scale. What is being built is not simply infrastructure, but a national capability.
The System Behind the Strategy
China’s computing power strategy offers a broader insight into the nature of competition in the digital age. The decisive factor may not be individual technologies, but the systems that enable them: systems that integrate infrastructure, energy, coordination, and markets into a coherent whole.
What China is constructing is a national architecture for organizing and mobilizing computing power. If successful, this system will not only support domestic development, but will also shape how other countries think about digital infrastructure and governance.
The competition, in this sense, is no longer just about innovation. It is about who can build the system that makes innovation possible.