The Zhitong Finance App learned that China Galaxy Securities released a research report saying that large-scale AI development is driving the rapid restructuring of infrastructure systems, and that collaboration between energy and computing power has become the key to supporting the development of new types of productivity. Stable, low-cost, and green energy supply capabilities will be an important factor affecting the layout and competitiveness of the AI industry. China is speeding up the construction of the “six networks”, promoting collaborative development of infrastructure such as energy, computing power, and communications, and building systematic support for industrial upgrading and economic growth in the AI era. Future large-scale construction of computing power infrastructure such as data centers and intelligent computing centers will not only mean an increase in demand for computing power equipment such as servers and chips, but will also simultaneously drive the expansion of infrastructure investment in power supplies, power grids, energy storage, and distribution systems. As far as the capital market is concerned, investment in AI computing power is gradually expanding from a simple “chip and server market” to a “power infrastructure market.”
The main views of China Galaxy Securities are as follows:
Artificial intelligence is driving a systematic restructuring of the energy demand structure and the global energy competition pattern. As computing power requirements shift from phased model training to large-scale inference services, data centers gradually evolved from peak demand to continuous and stable high baseload requirements, placing higher demands on the reliability, flexibility, and carrying capacity of power systems. AI data centers present new features such as high power density, high reliability, and strong spatial agglomeration. Energy constraints have expanded from simply meeting incremental power requirements to system capabilities such as power grids, energy storage, cooling, and infrastructure collaborative configuration. In the future, global AI competition will accelerate from competition for resource endowments to competition for energy-computing power collaboration, and a stable, low-cost, and green energy supply system will become the key to supporting the development of computing power and shaping the competitive advantage of the industry.
In the age of artificial intelligence, competition between China and the US is shifting from competition for energy resources to competition for energy system capabilities. Energy security capabilities, resource allocation capabilities, and cost control capabilities have become key factors determining the development of the AI industry. The US market has a flexible mechanism and strong short-term supply response capacity. China has a perfect infrastructure system and outstanding long-term coordination advantages, but both require further upgrading the modernization level of the energy system. The key to future AI competition is to compete for system capabilities that efficiently collaborate energy, electricity, computing power, and industrial applications.
Relying on a mature energy market, abundant natural gas resources, and nuclear power layout, the United States can quickly form reliable capacity and promote dynamic matching of energy and computing power through market mechanisms and long-term power purchase agreements. It has strong power supply guarantees and cost advantages in the short term, but it also faces slow grid expansion, insufficient transmission capacity across regions, and pressure for a low-carbon energy transition.
China, on the other hand, relies on major infrastructure projects such as the world's largest unified power grid, large-scale power systems, and “East Digital and Western Computing” to continuously improve reliable power supply capacity and energy-computing power collaborative allocation efficiency to provide systematic support for large-scale AI infrastructure construction. However, at the same time, the share of new energy generation in China continues to rise, bringing with it challenges such as insufficient flexible adjustment capacity of the power system and the need to enhance energy storage and cross-regional consumption capacity. The advantages of new energy resources have not yet been fully transformed into stable, low-cost computing power advantages.
Facing the AI era, energy systems need to shift from meeting demand for new electricity to building systems to support long-term development of computing power. It is difficult to meet AI's demand for large-scale, stable, low-cost, and low-carbon electricity over a long period of time by simply expanding power installations. Collaborative planning of energy and computing power should be promoted, computing power space allocation should be optimized according to energy resources, power grid carrying capacity, and industrial layout, and the conversion efficiency of energy resources into computing power capacity should be improved. At the same time, around computing power guarantee requirements under a high proportion of new energy sources, capacity building for energy storage, power grid regulation, and intelligent scheduling will be accelerated, power system flexibility will be enhanced, and AI will play the reverse enabling role of AI in new energy forecasting, power grid operation, and load regulation. On this basis, energy-computing power collaboration will be incorporated into the “six networks” construction, integrated planning with infrastructure such as communications and logistics, etc., to promote the collaborative layout of infrastructure networks with advanced manufacturing, strategic emerging industries, and future industries, and form a transformation chain of “energy resources - infrastructure - computing power - industrial competitiveness”.
Risk warning: 1. Risk of inadequate policy understanding; 2. Risk of policy implementation falling short of expectations; 3. Risk of uncertainty in technological development.