Five Metals Essential To The New Era Of AI Have Been Identified

• 19.08.2026, 13:37

Copper turned out to be the most important.

The rapid development of artificial intelligence requires not only more powerful chips, but also a vast amount of other resources. Building data centers requires power lines, transformers, cooling systems, servers, and networking equipment—and their production is impossible without metals, writes Interesting Engineering.

A study of data center infrastructure conducted in 2026 showed that copper accounts for about 83% of the total estimated mass of minerals required for such infrastructure. It is used in power supply systems, server wiring, transformers, and other data center components. Thanks to its high thermal conductivity, copper is also in high demand in cooling systems.
Gallium has become the second most important material. Although significantly less is required, it is used in the production of semiconductors and power electronics. The International Energy Agency classifies gallium as a raw material with a high risk of supply disruptions due to the concentration of production and the limited number of substitutes.

A similar situation exists with germanium. It is used in semiconductors and fiber-optic systems, which enable the transmission of enormous volumes of data.

Another important group consists of rare earth elements, including neodymium, praseodymium, dysprosium, and terbium. These are needed to produce powerful permanent magnets used in motors and various types of equipment. At the same time, China accounts for more than 90% of global rare-earth metal refining for magnets.

Aluminum rounds out the list. It is used in radiators and cooling systems due to its light weight, corrosion resistance, and good thermal conductivity.