Hebei, a traditional industrial powerhouse, is making a leap forward on the track of new quality productive forces, with the province's operational intelligent computing scale ranking first nationwide.

Recently, reporters ventured into the frontlines of Hebei's computing power industry to explore the path from "leading in scale" to "winning in efficiency." The core question: how can more value be created? The answer lies in converting computing power advantages into industrial strengths.

At the Runze Artificial Intelligence Application Center, the industry's first single-unit 200-megawatt intelligent computing center supports the training and inference of trillion-parameter models every day. Since settling in Langfang in 2009, Runze Technology has put 13 intelligent computing centers into operation, establishing itself as a leading enterprise in the national intelligent computing sector. This stands as a microcosm of Hebei's computing power industry development.

Leveraging its geographical advantages, Hebei has seized opportunities from the Beijing-Tianjin-Hebei coordinated development and the "East Data, West Computing" project. Targeting Beijing, the nation's largest computing power demand market, the province has accelerated the construction of a Beijing-surrounding intelligent computing cluster, with 88 large-scale and above computing centers built and operational.

But with scale advantage in hand, how does one create industrial value? In terms of intelligent computing scale, Hebei leads the nation. However, looking at the big data industry size, Hebei's 2025 big data industry revenue reached 34.76 billion yuan, ranking tenth nationwide.

Chinese Academy of Engineering academician Wu Hequan points out that every 1 yuan invested in computing power can drive 3 to 4 yuan of GDP growth. Despite Hebei's massive computing infrastructure, existing centers primarily rely on self-built, self-use models by internet companies and cabinet hosting operations. Most users are located outside the province, while local intelligent agent research and application lag behind. Exploration of new models like multi-agent collaboration and "large models plus intelligent agents" remains insufficient, meaning the computing power advantage has not yet been fully converted into industrial value and economic benefits.

During visits, reporters found that most of Hebei's computing centers play the role of "computing landlords." They rent out cabinets, supply electricity, and provide maintenance—low value-added services that offer limited support to the local AI industry. For example, the Heyun Data (Huailai) Technology Industrial Park serves almost exclusively out-of-province users, while the Financial Computing Center in Sanhe counts Beijing users as over 90 percent of its clientele.

Upgrading single-function computing infrastructure into diversified, high value-added service formats is a challenge Hebei's computing power industry must address. Runze Technology has already taken the lead. It delivered the nation's first whole-building liquid-cooled intelligent computing center and independently designed a new generation of intelligent computing centers. Seizing opportunities in the intelligent computing sector, Runze Technology has deployed nine AIDC (Artificial Intelligence Data Center) parks globally and is planning gigawatt-level intelligent computing parks to provide larger-scale, more advanced infrastructure services.

In the view of Zhou Chaonan, Chairman of Runze Technology, the computing power industry is undergoing profound changes: demand is shifting from general computing to intelligent computing, and onward to gigawatt-level AGIDC (next-generation intelligent computing centers for the general AI era). The value logic is pivoting toward tokens, and token production efficiency, usage cost, and service capability will directly enhance the core competitiveness of the intelligent computing industry.

In March of this year, China's average daily token invocation volume surpassed 140 trillion, growing more than a thousandfold in two years. In Hebei, cities like Zhangjiakou, Shijiazhuang, and Xiong'an New Area are accelerating the layout of token factories to drive the iterative upgrade of the computing power industry.

At the Changshan Cloud Data Center in Shijiazhuang, the token factory's invocation curve continues to climb. "Daily token invocations have exceeded 10 billion," says center director Pei Yong, who calculates: selling cabinets takes 10 years to recoup investment, selling computing power takes 5 years, but selling tokens achieves payback in under 2 years.

However, it must be acknowledged that Hebei is still in the early stages of extending the computing value chain. The province lacks sufficient capacity for large-scale self-owned computing and token supply nationwide, and there is a dearth of influential token operation platforms and service systems. In the intelligent era, building new forms of the intelligent economy requires accelerating the transformation toward "Token operations."

An Yun, Director of the Artificial Intelligence Research Institute at Sai Zhi Industrial Research, suggests Hebei should leverage its computing scale advantage to deploy ultra-large intelligent computing clusters and token production bases. This would push data centers and cloud service providers to further package underlying computing power into standardized products, creating a Beijing-Tianjin-Hebei coordinated Token production and operation system. This would enable computing power to be converted into AI services that can be called on demand and paid by usage, raising the service value per unit of computing power.

"The insufficient depth of intelligent agent integration with industries is the underlying reason for Hebei's weak computing power value conversion," says Guo Liang, Chief Engineer of the Cloud Computing and Digital Research Institute at the China Academy of Information and Communications Technology. Hebei is now promoting computing infrastructure as a competitive advantage for attracting investment.

Attracted by the computing power advantages and market prospects, global security market leader Zhejiang Uniview Technologies has set up its North China subsidiary in Langfang, cutting R&D costs by 40 percent and shortening model iteration cycles by 30 percent. The value of computing power is ultimately realized in application scenarios. Hebei boasts advantageous industries such as steel, equipment manufacturing, and pharmaceuticals—natural testing grounds for intelligent agents.

Wu Hequan believes Hebei's key move is to convert computing power advantages into scenario advantages, then translate scenario advantages into industrial advantages, thereby cultivating new economic growth points. "Computing power can spawn emerging industries such as data services and the intelligent agent economy. Hebei's leading computing scale nationwide will unleash enormous potential in the digital intelligence era," says Guo Liang.