Artificial Intelligence (AI) Competition, the US Finds Copper Is Not Enough

Super-large data centers are like industrial factories and require a large amount of reliable power. The International Energy Agency estimates that the electricity demand of global data centers this year will approach 1,000 terawatt-hours, equivalent to the energy consumption of the entire country of Japan. Such a huge demand means a significant demand for material resources, among which copper is key.

While many analyses focus on the energy or computing costs of ChatGPT searches, there is very little research on the material costs of such searches.

Copper is the fundamental metal for electrification, and the booming development of artificial intelligence adds a huge demand for copper on top of the expanding power system needs. The amount of copper required for data center racks in the AI era is two to four times that of traditional racks. Therefore, it is expected that the global demand for copper in data centers will more than double by 2030, and the cumulative copper consumption of the industry is expected to reach 4.3 million tons by 2035.

This surge in demand conflicts with an increasingly severe supply crisis. Analysts warn that structural gaps are imminent, potentially reaching 10 million tons by 2040. Supply is not flexible; it may take more than ten years to bring a new copper mine online globally, and as long as 29 years in the United States. To make matters worse, refining capacity is highly concentrated outside North America, and current market and pricing mechanisms are insufficient to effectively incentivize the construction of new infrastructure.

This resource competition conflicts with the supply chain of military systems. Communication equipment, electronic warfare systems, advanced radar, and hardened facilities all require large amounts of copper. When AI companies bid for power, purchase land near critical substations, or reach long-term supply agreements, they are fundamentally reshaping the industrial base, thereby lowering the production priority of military systems.

Source: The National Interest

Original: toutiao.com/article/1855884943250432/

Statement: This article represents the views of the author himself