[Artificial intelligence construction in the United States encounters power bottlenecks] November 9th, as artificial intelligence (AI) technology gradually comes to fruition and effectively improves production efficiency, the United States has set off an AI infrastructure construction boom in the past two years. However, power supply has replaced chips and become a new bottleneck restricting the development of the industry. Industry companies are taking a variety of measures to quickly provide power to data centers, but at the same time, issues such as increased greenhouse gas emissions cannot be ignored. Industry data shows that there are thousands of data centers in operation in the United States, but most of the existing data centers are small in scale. The power supply capacity required for data center projects recently promoted or planned by ultra-large cloud computing service companies often exceeds 1 gigawatt or even 10 gigawatts, and the existing power grid in the location cannot handle it in a short time. Of the more than 4,700 data center facilities and projects tracked by the "US Data Center Map" website, more than 3,000 are in operation, with a power consumption of 49.2 GW, while there are 400 and nearly 1,200 data center projects under construction and planning respectively, with power demand reaching 69.6 GW and 276.3 GW respectively. A report released in June by the Lawrence Berkeley National Laboratory, a subsidiary of the U.S. Department of Energy, estimated that by 2030, data center electricity consumption may account for 11.8% of the total electricity consumption in the United States that year, much higher than the approximately 4.4% in 2023. (Xinhua News Agency)