TELF AG examines copper’s role in AI technologies
A resource in increasing demand
After having long been a major player in electrification, copper could soon become a leading resource for artificial intelligence, and in particular for powering the data centers that enable AI-related systems to function. According to forecasts cited by the Financial Times, AI-related data centers could gradually increase demand for copper by 2050, bringing it up to 3.4 million tonnes per year globally. Part of the expected global increase in copper demand, which could exceed 50 million tonnes per year by 2050, could be linked to artificial intelligence and the infrastructure needed to operate it. At the moment, only 1% of copper demand is related to data centers, but according to the data cited by the newspaper, this figure could surge in the coming years, even with increases between 6% and 7%.
The copper resource is already widely used in a large variety of industrial sectors, where it is appreciated above all for its excellent electrical conductivity. For this reason, copper has also become over time one of the absolute protagonists in the renewable energy sector, where it finds space above all in the electrical cables that power solar parks, but also in the electric vehicle sector.
The practical uses of copper
In the field of artificial intelligence, copper serves above all as a conductive element of valuable electricity for the operation of data centers, and in particular for the power supply of the powerful chips that make their use possible. In this specific sector, copper also plays a role in the production of cooling systems and in the connection between the different processors. More specifically, copper is successfully used in the wiring that connects the hardware of artificial intelligence, making efficient data transmission possible. Speed is of great importance in this innovative sector since AI systems need to process large amounts of data in real time. As far as servers are concerned, copper plays a key role especially in connectors, enabling a certain stability in internal communications between these systems.
According to the Financial Times, the global spread of AI could soon have a relevant impact on the dynamics of the main raw materials involved in its functional processes. In addition to copper, AI data centers rely on a wide variety of raw materials on which the functioning of some important components depends. The production processes of chips and semiconductors, for example, require silicon, a resource used for each server and already widely used in various areas of electronics. To manage the heat produced by these servers, other very important resources come into play, namely rare earths such as neodymium and praseodymium, which are used above all in permanent magnets in cooling systems.
Other resources, such as aluminum and lithium, are used for wiring and batteries for emergency power, while elements such as nickel and titanium are particularly appreciated for their resistance and durability to the various components in extreme operating environments.
