2015至2024年间,美国数据中心并未推高居民平均电价,反而数据中心容量每增加10%,居民电价平均下降约0.4%。原因是电网建设与维护成本相对固定,大型稳定用户增加售电量后可摊薄单位成本,新增电厂运行成本也可能低于旧有电源。该效应成立的前提是电网能持续扩建足够供给,否则短缺或过度建设可能逆转结果。
Data centers did not raise average U.S. household electricity prices from 2015 to 2024.
Rather the opposite, their estimate finds: a 10% increase in data-center capacity lowered average residential electricity prices by about 0.4%.
Because, an electricity grid costs a lot to build and maintain whether it is lightly used or heavily used.
So when a large, steady customer like a data center buys much more electricity, those grid costs are spread across many more units of electricity sold. Total costs may rise, but the cost per unit can fall.
And when more power plants are needed, the new plants can be cheaper to run than the older power supply already built into customer bills.
The reason is not that electricity becomes free to produce, but that grids, wires, and generators can sometimes serve more demand at a lower average cost.
There is one important limit: this only holds when the grid can keep building enough supply, because shortages or overbuilding could flip the effect and raise prices.
Lambda's co-founder and CTO @stephenbalaban : offshoring the data-center buildout (out of USA) would lose construction/operations jobs plus the reindustrialization jobs that follow from grid upgrades (steel, aluminum, turbines, etc.) "AI is a national strategic asset, Compute is a national strategic asset." https://t.co/OXkpxsff15在 X 查看被引用的帖子
来源:@rohanpaul_ai · x.com