The rapid increase in AI data centers is happening at the same time as extreme heatwaves, which is putting a lot of pressure on the electricity grids and water resources of the United States.

AI 數據中心的快速增加正與極端熱浪同時發生，這給美國的電網與水資源帶來了巨大壓力。

The combination of higher computing needs and unusual heatwaves has caused a serious strain on national infrastructure. In the Mid-Atlantic region, PJM Interconnection has asked for federal permission to require data centers to use backup power during peak times to keep electricity stable for homes and businesses. Similarly, grid operators in Texas and Illinois have emphasized that the fast growth of these large facilities is a main cause of future reliability risks and higher costs for consumers. Furthermore, some developers use their own fossil fuel generators to avoid grid limits, but this practice often increases local environmental damage.

較高的運算需求與異常熱浪的結合，導致國家基礎設施承受嚴重壓力。在中大西洋地區，PJM Interconnection 已請求聯邦許可，要求數據中心在用電高峰期使用備用電源，以維持住家與企業的電力穩定。同樣地，德州與伊利諾州的電網營運商也強調，這些大型設施的快速增長是未來可靠性風險增加及消費者成本上升的主因。此外，部分開發商使用自有的化石燃料發電機以規避電網限制，但這種做法往往增加了對當地環境的破壞。

Water shortages are another serious problem. Data centers, especially those using evaporative cooling, use huge amounts of drinking water, and much of it is lost through evaporation. This is a major issue because most new centers are built in areas that already lack water. Additionally, there is a problem with corporate reporting; while companies like Meta, Google, and Amazon share some sustainability data, they do not provide clear information on indirect water use. This refers to the water used by power plants to create the electricity that runs the centers, which research shows can be much higher than the water used on-site.

水資源短缺是另一個嚴重問題。數據中心，尤其是使用蒸發冷卻的中心，會消耗大量飲用水，且其中許多水透過蒸發而流失。這是一個重大問題，因為大多數新中心都建在原本就缺水的地區。此外，企業報告也存在問題；雖然 Meta、Google 和 Amazon 等公司分享部分永續發展數據，但並未提供關於間接用水的明確資訊。這指的是電廠為運行數據中心而發電時所使用的水，研究顯示這類用水量可能遠高於場地內的使用量。

There is a growing disagreement between industrial goals and public interests. While the federal government has generally supported this growth, some regional leaders have taken a stricter approach. For example, Texas Governor Greg Abbott has argued that rural data centers should be banned unless they can provide their own resources. At the same time, some federal lawmakers have proposed stopping new developments. Public opinion is also critical, as many people oppose these projects due to resource depletion. To solve these issues, experts suggest using closed-loop cooling systems, geothermal energy, and more efficient AI models to reduce the total energy needed.

工業目標與公眾利益之間的分歧日益增加。雖然聯邦政府大致支持這種增長，但部分地區領導人採取了更嚴格的做法。例如，德州州長 Greg Abbott 主張，除非鄉村數據中心能提供自有資源，否則應予以禁止。與此同時，部分聯邦立法者建議停止新開發項目。公眾輿論也至關重要，許多人因資源枯竭而反對這些項目。為了解決這些問題，專家建議使用閉環冷卻系統、地熱能以及更高效的 AI 模型，以降低總能源需求。

The United States is currently facing a difficult balance between expanding its AI capabilities and the physical limits of its energy and water systems.

美國目前在擴展 AI 能力與能源及水務系統的物理限制之間，面臨著艱難的平衡。