The global technology sector is currently expanding its data center infrastructure at an unprecedented rate to support artificial intelligence. This growth involves spending huge amounts of money and carefully choosing different locations around the world to build these facilities.

全球科技產業目前正以前所未有的速度擴展數據中心基礎設施，以支持人工智慧。這種增長涉及投入巨額資金，並在全球不同地點 carefully 選擇地點來興建這些設施。

The growth of AI infrastructure is clearly seen in the Nscale project in Narvik, Norway. This facility, built for Microsoft and OpenAI, uses the region's extra hydroelectric power and cold climate to keep cooling costs low. However, corporate strategies are changing; for example, OpenAI has shifted from being a main partner to renting capacity, which suggests they are trying to be more financially careful before going public. While Nscale has received significant funding, including $2 billion and investments from Nvidia, some analysts warn that the company relies too heavily on debt and assumes that the demand for computing power will stay high.

AI 基礎設施的增長在挪威納爾維克的 Nscale 項目中清晰可見。這個為微軟與 OpenAI 興建的設施，利用該地區多餘的水力發電與寒冷氣候，將冷卻成本降低。然而，企業策略正在改變；例如 OpenAI 已從主要合作夥伴轉為租用容量，這表明他們在上市前試圖在財務上更加謹慎。雖然 Nscale 獲得了大量資金，包括 20 億美元以及來自 Nvidia 的投資，但部分分析師警告，該公司過於依賴債務，且假設對運算能力的需求將持續高漲。

At the same time, the industry is facing resource shortages and public protests. In the United States, especially in Ohio, people are worried about the environmental impact of these data centers. The main concerns are the loss of local water supplies used for cooling and the possibility that electricity prices will rise for regular consumers. Furthermore, some economists argue that government tax breaks for these companies are not effective, leading to new laws to remove these subsidies. To solve these problems, Google is implementing water replenishment programs, while other companies are stopping the use of water-based cooling entirely.

與此同時，產業正對面資源短缺與公眾抗議。在美國，特別是在俄亥俄州，民眾擔心這些數據中心對環境的影響。主要擔憂在於用於冷卻的當地水源流失，以及普通消費者電價上漲的可能性。此外，部分經濟學家認為政府給予這些公司的稅務減免並不有效，導致出台新法規以取消這些補貼。為了縮小這些問題，Google 正在實施水資源回補計畫，而其他公司則完全停止使用水冷系統。

Because of the urgent need for speed, some companies are using unusual construction methods. Meta has used 'rapid deployment structures'—which are essentially large, weatherproof tents—in Ohio to install AI chips faster than traditional buildings allow. These sites often use modular gas turbines to ensure they have reliable power. This move toward fossil fuels, also seen in Nscale's West Virginia operations, shows a conflict between the immediate need for power and long-term environmental goals, as renewable energy grids cannot keep up with the fast pace of AI development.

由於對速度的急切需求，部分公司採取了不尋常的建築方式。Meta 在俄亥俄州使用了「快速部署結構」——在本質上是大型的防風雨帳篷——以便比傳統建築更快地安裝 AI 晶片。這些場地通常使用模組化燃氣輪機以確保電力可靠。這種向化石燃料轉向的趨勢，在 Nscale 位於西維吉尼亞州的營運中亦可見，顯示出電力即時需求與長期環境目標之間的衝突，因為再生能源電網無法跟上 AI 發展的快節奏。

The race to build AI infrastructure continues to speed up because of the need for computing power and cheap energy. However, this growth is facing increasing challenges from government regulations, environmental concerns, and social opposition.

由於對運算能力與廉價能源的需求，建設 AI 基礎設施的競賽持續加速。然而，這種增長正面對來自政府監管、環境憂慮以及社會反對的日益增加之挑戰。