Artificial intelligence (AI) is often described as a digital revolution driven by algorithms and data. In reality, it is deeply rooted in a complex physical and industrial ecosystem. The illusion of the “cloud” has fueled the idea of an immaterial technology, but data centers are heavy industrial infrastructures that require land, servers, cooling systems, and redundant power supply. AI relies on massive, energy-intensive, and geopolitically concentrated infrastructures. Far from being immaterial, it represents one of the most resource-intensive technological transformations of the 21st century.
The explosive growth of generative AI clearly demonstrates this. ChatGPT surpassed one billion users in just three years, highlighting an unprecedented compression of technology adoption cycles. This acceleration demands a rapid expansion of computational capacity and a growing demand for energy, metals, and minerals.
Data centers are large-scale hubs that collect, store, and process enormous volumes of data, requiring extensive network infrastructures and vast amounts of electricity. Copper, thanks to its high conductivity, is indispensable in electrical systems and therefore crucial for data centers. At the core of AI systems are advanced semiconductors, particularly GPUs and specialized chips, whose production requires critical materials, rare earths, and industrial gases, increasing the complexity of supply chains.
MINERALS AND ENERGY FOR DATA CENTERS
AI infrastructure requires a wide range of critical minerals, while the ability to secure, process, and integrate these resources becomes a determining factor for technological leadership and geopolitical influence.
The installation of data centers requires stable and always available energy. Not surprisingly, Mark Zuckerberg stated that “energy, not computing power, will be the main bottleneck for AI progress.” Industrial metals are therefore central also in the electrification dynamics that support AI expansion.
Each unit of energy used for computing generates significant heat. Consequently, data centers consume enormous amounts of electricity, both for processing and cooling. This creates an indirect dependence on natural gas, nuclear, and renewables. Today data centers account for about 2% of global electricity consumption, a share that could rise to 4% by 2035.
AI BETWEEN INNOVATION AND GEOPOLITICS
AI is no longer just an innovation tool but has become a factor of economic, scientific, and military dominance, redefining global geopolitical balances. At the heart of this transformation are semiconductors, with the rivalry between the United States and China crystallizing into what has been called the “semiconductor war.”
Strategic supremacy in AI is driving massive investments: the United States has invested 335 billion dollars over ten years, while China has significantly increased its presence. AI thus becomes the main arena of global competition.
At the same time, structural limits emerge: developing new mines takes a very long time and only 70% of copper demand can be met by 2035.
China dominates the entire metals value chain, creating a strong dependence of Western economies and acquiring significant strategic leverage.
In this context, industrial commodity prices are destined to maintain a structurally upward trend, supported by growing demand and limited supply. In the short term, however, volatility will remain high.
For this reason, governments and companies are working to diversify supply chains and relocate production, increasing investments in infrastructure. For investors, this scenario opens opportunities along the entire supply chain, from mining activities to refining and recycling.
