On March 31, at the Hannover Industrial Fair in Germany, a visitor threw objects in a virtual world through VR equipment. Photo by Xinhua News Agency reporter Zhang Fan

  In recent years, artificial intelligence has developed rapidly, not only has it been used more and more widely in the energy field, but it is also expected to bring huge changes to the global energy industry in the next 10 years. On the one hand, with the continuous construction of global data centers supporting artificial intelligence infrastructure, global power demand will further increase; on the other hand, the rapid development of artificial intelligence will also create opportunities for reducing production costs, improving product competitiveness, and reducing greenhouse gas emissions. Affordable and sustainable power supply is a key factor in the development of artificial intelligence and puts the energy industry at the forefront of today's most important technological revolution.

  The research and development and application of artificial intelligence requires a large amount of high-energy-consuming data center support. Since 2022, various countries and regions have almost doubled their investment in data centers, and the total investment in 2024 has reached US$500 billion. The International Energy Agency recently released a special report on "Energy and Artificial Intelligence" showing that large-scale data center construction has promoted a significant increase in global electricity demand. In 2024, data centers account for about 1.5% of global electricity consumption, reaching about 415 terawatt hours. Among them, the United States accounts for the largest proportion of global data center power consumption, reaching 45%, and Europe accounts for about 15%. The report shows that in the past five years, the proportion of global data centers in global electricity consumption has increased by 12% per year. At the current rate, global data center power demand will more than double to about 945 terawatt hours by 2030. By 2030, the demand for electricity consumption in data centers serving artificial intelligence will increase by more than 4 times. By 2035, global data center power consumption will climb to about 1,200 terawatt hours.

  The proportion of electricity consumption in data centers in developed economies around the world will increase significantly. At the beginning of 2025, the US government announced that it would spend at least US$20 billion to build data centers in its own country. The EU also announced the launch of the "Invest AI Plan" to provide a special fund of 20 billion euros. Japan actively cooperated with multinational technology companies such as Google, Microsoft, and OpenAI to expand the scale of data center construction in its own country. Correspondingly, from 2024 to 2030, the increase in electricity consumption of data centers in the United States will account for nearly 50% of the country's total electricity demand increase. By 2030, the electricity consumption used in artificial intelligence data processing in the United States will exceed the total electricity consumption of traditional energy-intensive industries such as aluminum, steel, cement and chemicals; the increase in electricity consumption of Japanese data centers in their total national electricity demand growth will exceed 50%; the electricity consumption of EU data centers will reach 150 terawatt hours by 2030, an increase of about three times from the current scale. In contrast, by 2030, the increase in electricity consumption in data centers in developing countries and emerging economies will only account for about 5% of the total increase.

  At present, all countries are taking diversified energy methods to meet the surge in power demand for data centers based on their own actual conditions. Among them, renewable energy and natural gas are expected to become the main force in meeting the power demand of data centers with its cost advantages and a mature supply system. It is estimated that by 2035, renewable energy generation will increase by more than 450 terawatt hours, natural gas generation will increase by 175 terawatt hours, and more than 50% of the new electricity demand in global data centers can be met through the above two energy sources. In addition, with the first batch of small modular nuclear reactors expected to be put into operation around 2030, nuclear energy will also become an important support for meeting the power needs of data centers.

  Such rapid and substantial growth in demand also poses challenges to the global energy sector. First of all, power grid systems in many countries and regions are approaching their load bearing limits. According to the International Energy Agency, if the power industry is not upgraded at accelerating, about 20% of planned data center projects around the world may face the risk of delay. Secondly, there are many uncertainties in the speed of popularization and efficiency improvement of artificial intelligence. Only 12% of oil and gas fields in the world have completed digital transformation, the coverage rate of smart grids in developing countries is less than 40%, and the large gap in compound talents has led to the penetration rate of artificial intelligence technology in the energy industry being lower than that of other industries. Again, artificial intelligence and the energy industry lack integration. The commercialization of new technologies in the energy field often takes decades, and it requires increasing R&D investment to shorten the business transformation cycle, but among current energy science and technology enterprises, only 2% of equity financing flows to artificial intelligence companies. At the same time, most energy companies are also facing multiple constraints on digital security, supply chain security and physical security risks. Especially the current arbitrary tariff policy of the US government has also pushed the key mineral supply chains closely related to artificial intelligence into dangerous situations.

  The impact of the application of artificial intelligence in the energy field on climate change is also one of the global focus. The International Energy Agency expects global data center carbon emissions to climb from the current 180 million tons to 300 million tons by 2035. Although the total emissions are less than 1.5% of the total emissions of the energy industry, data centers have become one of the fastest-growing sources of emissions. It is worth noting that the emission reduction benefits that can be generated by the large-scale application of artificial intelligence technology far exceed the data center's own emissions. By 2035, if existing artificial intelligence solutions can be promoted and applied, the emission reduction potential is about twice that of Japan's current annual emissions. Data center construction can also become a lever for energy transformation, and a single super-large data center can leverage the construction of surrounding 300-megawatt low-carbon power clusters.

  Fatih Birol, Director of the International Energy Agency, said, "With the rise of artificial intelligence, the energy industry is standing at the forefront of the most important technological revolution of our era." Although the future is still full of uncertainty, with the policy support and strong promotion of governments, the technology research and strategic coordination of the two major industries has been launched, and the global energy landscape and the global energy industry are being reshaped. (Reporter Wang Baokun)

[Editor in charge: Ran Xiaoning]