In recent months, the demand for AI services powered by DeepSeek has surged, prompting major internet companies and the three largest telecom operators in China to fast-track their investments and construction of cloud and AI hardware infrastructure. Policymakers across various regions and departments are actively supporting this initiative through financial investments and policy measures aimed at fostering advancements and innovation in key technologies related to cloud computing. Industry analysts predict that the advent of large-scale AI models will spearhead a new growth phase for the cloud computing industry, maintaining a vibrant market in China. Projections estimate that by 2027, the Chinese cloud computing market could exceed 2.1 trillion yuan.

On February 24, Alibaba Group announced its commitment to investing more than 380 billion yuan over the next three years for enhancing its cloud and AI hardware infrastructure—surpassing the total investment made over the past decade. This investment marks the largest endeavor by a private enterprise in China to bolster its cloud and AI hardware capabilities. Alibaba Group’s CEO, Wu Yongming, emphasized the extraordinary growth of AI and the immense potential it presents for the domestic tech industry, stating that the company is fully dedicated to accelerating the construction of cloud and AI infrastructure to foster ecological development across various sectors.

Moreover, the three major telecom operators have also outlined their strategies for building cloud infrastructure, fast-tracking investments in computing centers and related projects. Recently, a major project signing event took place in the Lin-gang New Area of Shanghai, which included plans for China Telecom's Lin-gang Information Park. This enormous project aims to establish a super computing center capable of supporting training for large AI models that require trillions of parameters. Meanwhile, China Unicom has initiated construction on its Beijing-Tianjin-Hebei Digital Technology Industrial Park (Northern District), which is projected to reach significant construction milestones by July 2025, further promoting the development of computing data platforms to enhance industrial upgrades. Additionally, in its 2025 work conference, China Mobile stressed the critical nature of infrastructure construction and the importance of optimizing data center layouts to maximize the benefits of AI and scalability.

Industry experts have noted that the frequent appearances of “server busy” messages from DeepSeek are indicative of the growing demand exhausting available computing power and congesting networks. Skyrocketing user demand for AI services, fueled by the proliferation of DeepSeek models, is expected to lead to an exponential increase in overall AI service requirements. This demand subsequently catalyzes rapid advancements in AI computing power.

A recent report released by the International Data Corporation (IDC) and Inspur Information projected that by 2025, the scale of intelligent computing power in China would reach an astonishing 1,037.3 EFLOPS (a measurement indicating quintillions of floating-point operations per second), representing a 43% increase compared to the previous year. Furthermore, by 2026, it is anticipated that this figure will double to 1,460.3 EFLOPS.

As stated by IDC’s Vice President Zhou Zhenggang, while the improvements in algorithm efficiency brought about by DeepSeek might appear to lessen the demand for computing capabilities, the reality is that an influx of new users and applications is driving the broader adoption and application of large models. This not only reshapes the paradigm of industry innovation but significantly enhances the construction of data centers, edge computing, and endpoint computing.

Simultaneously, the generative AI boom and the rising demand for large-scale models have shaped a new growth phase for the computing power industry. By 2027, China's cloud computing market is poised to surpass the estimated 2.1 trillion yuan.

According to Li Wei, Deputy Director of the Cloud Computing and Big Data Research Institute at the China Academy of Information and Communications Technology (CAICT), the disruption caused by generative AI and large models has created new opportunities for the cloud computing market. Mid-tier vendors, leveraging the strengths of technological platforms, can capitalize on this opportunity to accelerate the intelligent transformation of existing cloud businesses and gain market share, potentially reshaping the competitive landscape in the process.

Recently, multiple government departments have intensified their policy deployments to solidify the foundation of computing power, placing significant emphasis on advancing research and innovation in key technologies related to cloud computing.

The Ministry of Industry and Information Technology (MIIT) has issued a notice to organize actions aimed at establishing a robust computing power foundation. The focal points include six key areas of computation, storage, networking, application, environmental sustainability, and security in the computing network. The initiative seeks to discover enterprises and institutions with core technologies and strong innovative capabilities, fostering breakthroughs in several landmark technological products and solutions. The MIIT has committed to coordinating various resources to support high-performing entities, thereby promoting the demonstration and application of outstanding results. The recent call to fortify the foundation of computing power is intended to provide robust support for technological breakthroughs and practical applications.

In addition, many regions have been ramping up efforts to reinforce the construction of computing power infrastructure and industry application through financial support and robust policy measures. For instance, Hunan province has recently introduced various policy initiatives to foster high-quality development in the green intelligent computing industry. This includes support for major intelligent computing infrastructure projects with a scale exceeding 1,000 PFlops on a case-by-case basis, as well as financial assistance for projects with a total computing power above 50 PFlops, provided that the energy utilization efficiency of the data centers does not exceed 1.3, covering up to 10% of the total construction costs, capped at 20 million yuan.

Looking ahead, Li Wei suggested strengthening the collaborative efforts to achieve technological breakthroughs and maintaining a balanced supply-demand relationship for essential resources across multiple levels. He advocated for the construction of additional cloud infrastructure nodes, such as intelligent computing centers and supercomputing centers, while also minimizing losses associated with existing solutions and optimizing the effectiveness of current computing and cloud storage resources. Enhanced digitization across various scenarios is essential to support AI applications’ needs for intelligent computing power and empower diverse sectors of the economy.