In response to the burgeoning demands of AI applications, cloud service providers are ramping up their infrastructure deployment efforts. It is forecasted that by 2026, the global power demand capacity for data centers will soar to 161GW, marking a year-on-year increase of approximately 31%. AI servers are anticipated to be the main catalyst for this growth, contributing a significant 33.4%. The growth trajectory is expected to sustain a rate above 30% in 2027 as well. Nevertheless, beginning in 2026, a disparity between power supply and demand in data centers will start to become apparent, with the power gap widening considerably after 2028. By 2030, power demand is projected to skyrocket to 490.7GW, whereas the available grid capacity will only reach 222.6GW, leading to a staggering deficit of 268GW. This gap should be analyzed from two angles: firstly, manufacturers' self-built power generation facilities can partially bridge the gap, thereby reducing the actual shortage compared to the apparent disparity; secondly, delays in grid interconnection and transmission/distribution infrastructure construction will exacerbate the power supply shortage.