In the era of intelligence, data flows abundantly. As projects like "East Data and West Computing" enable the computing power network to comprehensively cover Digital China, storage capacity emerges as the linchpin for the seamless flow of data elements and AI innovation. Only through the harmonious development of storage and computing can true AI productivity be cultivated.

Coinciding with the 2025 World Artificial Intelligence Conference (WAIC), China Mobile and Sugon Storage jointly unveiled the nation's first intelligent storage capacity scheduling platform – a globally unified file storage system for computing centers. Positioned as the "central nervous system" of advanced storage capacity centers, this platform deeply integrates heterogeneous resource scheduling with AI application scenario requirements, offering core support for the efficient operation of advanced storage capacity centers and the implementation of AI technology.

This marks a pivotal shift in AI infrastructure, transitioning from a "computing-power-dominated" approach to a holistic "computing and storage" planning strategy. As He Zhen, President of Sugon Storage, remarked, "The cross-domain flow of data and AI innovation will become the norm, necessitating continuous efforts to overcome the underlying technical challenges of storage and computing collaboration. The launch of the intelligent storage capacity scheduling platform establishes a robust model for the coordinated development of storage and computing."

In recent years, the construction of AI infrastructure has gained momentum. However, the primacy of computing power has led to a phenomenon where "computing power is emphasized while storage capacity is neglected," resulting in storage capacity development lagging behind that of computing power.

In reality, storage capacity refers to the comprehensive ability to process data elements, tasked with providing a steady stream of "production materials" for AI and other application scenarios. With the rapid evolution of AI technologies, such as large models, the rise of heterogeneous computing power, and the increasing demand for massive data, the importance of storage and computing collaboration and holistic planning has accelerated.

Therefore, the collaboration between storage and computing urgently needs to enter a new phase. The nation's first intelligent storage capacity scheduling platform, jointly launched by China Mobile and Sugon Storage, sets an example for the coordinated development of storage and computing, offering high reference value for the future construction, operation, and application of AI infrastructure.

Reportedly, the intelligent storage capacity scheduling platform has taken the lead in covering the four national hub nodes of the Yangtze River Delta, Chengdu-Chongqing, Inner Mongolia, and Guizhou, spanning seven storage capacity resource pools across the east and west, and connecting the north and south. Furthermore, the localization rate of storage chips exceeds 90%, compatible with 8 types of AI chips worldwide, efficiently supporting the needs of AI technology implementation and AI application scenario deployment.

"Under the strategy of 'East Data and West Computing,' a unified computing power scheduling platform has emerged. Now, the launch of the intelligent storage capacity scheduling platform marks a crucial step for operators in implementing 'East Data and West Computing' and fostering the coordinated development of storage and computing," said Zhang Xinfeng, Vice President of Sugon Storage.

As we all know, computing power is crucial for the advancement of AI technologies, such as large models. However, it is high-quality massive datasets that truly drive qualitative changes in these models. It can be said that without a solid, reliable, and advanced intelligent storage capacity center, the full realization of data value is impossible, and no matter how powerful the computing power is, it will be in vain.

If the advanced storage capacity center is akin to a core brain that activates the value of data elements and nurtures AI productivity, then the intelligent storage capacity scheduling platform functions as the "central nervous system" of this brain. It is expected to thoroughly address persistent issues such as "incomplete visibility, disorganized management, and inefficient utilization of data," ushering in a leap in the efficiency of storage and computing collaboration.

Yang Zhilei, Vice President of Sugon Storage, stated, "Performance and efficiency are the core challenges of AI infrastructure. Among them, the challenge of data processing will become increasingly significant, necessitating more efficient data processing to achieve better storage and computing collaboration."

Reportedly, the intelligent storage capacity scheduling platform is a vital platform for executing the "East Data and West Computing" strategy. By uniformly managing mainstream storage products, intelligently analyzing and managing massive metadata, it realizes data value recognition of file systems, automatic migration scheduling across wide area networks, and business-agnostic access. The platform truly breaks down "resource islands" through "global scheduling" and promotes "coordinated computing and storage" through "data flow." Its value lies not only in single-point breakthroughs, such as accelerating large model training and real-time inference, but also in providing foundational support for the large-scale implementation of the AI industry by establishing a green, intensive, open, and inclusive storage capacity ecosystem.

Specifically, the intelligent storage capacity scheduling platform first enables data elements to truly flow. The platform accesses storage products from multiple vendors and regions through standardized interfaces, enabling unified monitoring and management of distributed storage, all-flash arrays, object storage, etc., completely eliminating data islands. This lays a solid foundation for the construction of high-quality datasets and effectively supports tasks such as large model training and inference.

For instance, the intelligent storage capacity scheduling platform supports P2P data transmission across clusters and domains, solving the problem of latency and improving data flow efficiency by more than 10 times.

Secondly, the intelligent storage capacity scheduling platform enables AI infrastructure to be truly efficiently utilized. By comprehensively analyzing business workloads and data hotness and coldness, it realizes intelligent migration of data grades, enabling efficient utilization of storage resources within the computing power network. Simultaneously, it adapts to various AI computing power chips, fully enhancing the efficiency of storage and computing collaboration, and ensuring AI clusters are "fully utilized."

Thirdly, the intelligent storage capacity scheduling platform makes the integration and innovation of AI and scenarios truly feasible. The platform enables the adaptation of multimodal AI scenarios, including autonomous driving, biomedicine, AIGC creation, etc., effectively supporting the data storage needs of various AI scenarios.

"The core capabilities of the intelligent storage capacity scheduling platform are dynamic allocation of infrastructure resources, integration of heterogeneous resources, and seamless cross-domain data transmission. Currently, the intelligent storage capacity scheduling platform is officially launched in the operator field and will have extensive applications in cloud service providers, industry leaders, and more in the future," Zhang Xinfeng concluded.

Inclusive AI is the core mission of building Digital China.

Among them, advanced storage capacity plays a crucial role. With the launch of the intelligent storage capacity scheduling platform, the future of advanced storage capacity looks even more promising. Judging from the actual changes in market demand, after the completion of the "East Data and West Computing" construction period, how to fully leverage AI infrastructure is of great significance for inclusive AI, and advanced storage capacity will inevitably play an even greater role in this process.

On the other hand, inclusive AI implies further lowering the threshold for using AI, enabling a vast number of small and medium-sized enterprises to conveniently adopt AI technology. Through the construction of advanced storage capacity, cloud service providers are expected to bring about the inclusiveness of computing and storage power, as well as the openness of data interfaces, which is conducive to fully harnessing the value of AI infrastructure and promoting different types of enterprises to embrace AI trends.

"Large models, autonomous driving, embodied intelligence, drug discovery... AI scenario demands are exploding, and inclusive AI is of great significance to industry digitization and intelligence. Sugon Storage will continue to adhere to independent research and development, continuously delve into fields such as AI storage and advanced storage capacity, and jointly accelerate AI application innovation with the industry," He Zhen concluded.