Self-Sufficiency and Security: Analysis of China's Smart Computing Development Strategy and Its Technological Limitations

The current phase of global scientific and industrial transformation positions Smart Computing (Zhìsuàn, 智算) as the fundamental productive force and a critical element of China's national security. The explosion in computing demand, particularly driven by generative artificial intelligence, has made technological self-reliance and strength a top priority. As President Xi Jinping stated, "A nation thrives through scientific advancement; a country grows strong with technological prowess." 

The Chinese strategy, informed by technical reports from institutions like China Academy of Information and Communications Technology (CAICT - 中国信通院), China Petroleum, and Mobile Cloud in 2025, is a coordinated plan to overcome supply chain vulnerabilities. The success of Chinese modernization, according to Xi, "hinges on the support of scientific and technological modernization". The analysis shows that State-Owned Enterprises (SOEs) are the driving force behind the construction of this critical infrastructure.

GettyImages

I. Political and Strategic Scenario: The Race for Technological Command

The Chinese central government has established the clear strategic objective of "accelerating the building of a country of innovators" and "becoming a country strong in science and technology by 2035." This ambition is driven by the realization that "the technological revolution and major-country competition are increasingly intertwined," making "the high-tech sector... both the forefront and main arena of international competition".

A. The Imperative of Self-Reliance and Strength

President Xi Jinping has elevated technological security to an existential issue, emphasizing the need to "maintain a firm grasp on the lifelines of science and technology and the initiative in development."

• Centralized Planning. Since the 18th CPC National Congress in 2012, an "innovation-driven development strategy" has been steadily promoted. The leadership of the CPC Central Committee over science and technology work is the foundation, ensuring that science and technology "always remain on the right development track."

• Strategic Objective:.The goal is to overcome the constraint that "certain core technologies in key areas remain constrained by external factors."The country must "intensify efforts to advance innovation and strive to secure the strategic high ground in global technological competition and future development."

  
  

B. The “Eastern Data, Western Computing” Project (东数西算)

The Smart Computing strategy is implemented through this key infrastructure project to optimize the distribution of computing resources.

• Deployment. All 8 national computing hub nodes have commenced construction.

• Strategic Layout:

  • Eastern Hubs: Focused on latency-sensitive workloads (e.g., Industrial Internet, AI Inference, financial securities).

  • Western Hubs. Focused on compute-intensive workloads (e.g., Large Model Training, offline analysis) and leveraging lower energy costs. This layout promotes "Green Computing" (Lǜsè Suànlì).

• Role of SOEs. Companies like China Mobile, China Unicom, and China Telecom are leading the rollout. China Mobile is implementing an "N+X" network layout, which substantially covers the East-West hubs.

II. Internal Technological Development and the Smart Computing Platform

The acceleration of AI, where training models like Sora may require $170+$ times the computing power of GPT-3 is driving domestic innovation.

A. Technological Autonomy: Chips and Expanded Architecture

The drive for technological self-reliance focuses on optimizing existing hardware and creating a software stack that mitigates reliance on high-end foreign chips²³.

• Production Capacity. China has achieved mass production of chips at $28 \text{ nm}$ and $14 \text{ nm}$. Companies like SMIC (中芯国际) are accelerating R&D on the $7 \text{ nm}$ process using DUV technology.

• Overcoming Bottlenecks. The strategy requires "intensifying R&D efforts with a focus on key areas and weak links." This means removing bottlenecks in areas such as "integrated circuits, industrial machine tools, basic software, advanced materials, research equipment, and core germplasm."

• Heterogeneous Management (Pooling). The critical technology is Heterogeneous Computing Power Resource Pooling, which aims to unify CPU, GPU, FPGA, and NPU/ASIC resources.

  • Software Abstraction. Techniques like function library layer virtualization (API interception) are used to manage different programming interfaces, such such as Huawei CANN, NVIDIA CUDA, and AMD ROCm, facilitating unified scheduling.

  • Interconnection. High-speed interconnection is supported by domestic protocols such as Huawei UB, Alibaba ALS, Tencent ETH-X, and China Mobile OSIA.

    
    

B. Scheduling Platform and Construction Models

Efficient resource management is delegated to the Smart Computing Platform, a five-layer architecture (Hardware, Resource, Framework, Operation, Application).

• Standardization and Scheduling: Kubernetes (K8s) has become the de facto standard for orchestration Platforms like Zhenze MaaS of China Mobile use Kosmos distributed cloud-native technology to support cross-domain, cross-chip hybrid training. China Mobile manages a cluster of over $100,000$ GPU cards and a total Smart Computing scale of $61.3 \text{ EFLOPS}$.

• Construction Models (SOE Focus). SOEs prioritize Self-Developed (自研) models for core business systems requiring high security and stability. Examples include China Construction Bank's "Jianxing Cloud" and China Petroleum's Kunlun model.

  
  

C. Outcomes and Industrial Transformation

Smart Computing is driving growth in emerging industries (integrated circuits, AI) and the transformation of traditional sectors.

• Energy. China Petroleum has developed the large Kunlun model, including a language model (300 billion parameters) and a multimodal model (80 billion parameters).

• Finance. China Construction Bank saw its non-CPU computing capacity exceed $23.39\%$ by the end of 2024. The bank is constructing a mega-center in Horinger New District (Inner Mongolia) with a planned investment of 10 billion RMB.

• Smart Cities. The scheduling platform of Shenzhen Zhicheng Yiyun (深圳市智城翼云科技有限公司) unifies over 140 resource pools and aims to provide $40,000 \text{ PFLOPS}$ by the end of 2025.

  
  

III. Strengths and Weaknesses of the Chinese Strategic Model

The Chinese strategy is a centrally coordinated action aimed at achieving digital autonomy.

A. Strengths

• Hierarchical and Comprehensive Planning. The CPC Central Committee ensures "centralized, unified leadership."

• National Mobilization. China leverages "the strengths of China's socialist system to pool resources for major undertakings."

• Institutional Integration. Efforts are underway to integrate "education, science and technology, and talent."

  
  

B. Weaknesses and Challenges

• Limited Original Innovation Capacity. President Xi acknowledged that "our capacity for original innovation remains relatively limited."

• External Technological Constraints. The problem that "certain core technologies in key areas remain constrained by external factors" persists.

• Talent Shortage: There is a "shortage of top-tier scientific and technological talent" and "pronounced structural mismatches between the available talent pool and the demands of scientific and technological innovation." The priority is to "accelerate efforts to build a contingent of professionals with strategically critical expertise."

Conclusion

China's strategic pursuit of dominance in Smart Computing (Zhìsuàn) is a critical national mission, seamlessly blending ambitious technological goals with geopolitical imperatives^. As articulated by President Xi Jinping, the endeavor to build a country strong in science and technology by 2035 is foundational to national rejuvenation and modernization.

The core of this strategy is achieving technological self-reliance and strength^. This focus is driven by the reality that "the high-tech sector has emerged as both the forefront and main arena of international competition" and that "certain core technologies in key areas remain constrained by external factors." The response is multi-faceted:

• Infrastructure Shift. The "Eastern Data, Western Computing" project is restructuring the national computing landscape, optimizing resource allocation between latency-sensitive eastern regions and energy-rich western hubs.

• Domestic Innovation. Efforts are intensifying to overcome bottlenecks in key areas like integrated circuits, basic software, and advanced materials. While China faces a significant challenge in advanced lithography (currently mass-producing at 28 nm and 14 nm), the focus is on maximizing the efficiency of domestic hardware through unified resource pooling and homegrown interconnection protocols.

• Institutional Strength: The system leverages the CPC's unified leadership to mobilize national resources effectively. State-Owned Enterprises (SOEs) lead the charge, rapidly deploying vast computing centers (such as the China Construction Bank's 10 billion RMB project) and self-developing platforms for core business systems, ensuring security and autonomy.

In essence, China is integrating its national development and security objectives under the banner of scientific innovation. While facing ongoing hurdles in original innovation capacity and a shortage of top-tier talent, the coordinated, top-down approach demonstrates unwavering strategic resolve to secure the nation's technological future and global influence.