Even AI is affected by the current energy crisis: data centers at risk in the United States

The AI sector is constantly growing, as we know. The large companies in the sector are now among the most highly capitalized in the world. For example, Nvidia shares have recorded a strong increase over the year, exceeding 150 dollars, with quarterly revenues of 38 billion dollars, up 92% compared to the previous year.

However, not all that glitters is gold. In the United States, the situation related to data centers, the real backbone of artificial intelligence, could be compromised: between 30% and 50% of projects planned for 2026 are likely to be canceled or severely delayed.

The main problem is not demand, which continues to grow explosively, nor a lack of investment, especially from the major technology giants. Google, Amazon, Microsoft and Meta are in fact allocating hundreds of billions of dollars to expanding their infrastructures. The real obstacle is physical and industrial: building data centers today is much more complex than it was even a few years ago.

Even in this context, the geopolitical situation and the energy crisis do not help. Data centers require enormous amounts of energy. As reported by ZeroHedge, the rapid expansion of artificial intelligence is facing increasingly clear limits in the ability of the global industrial system to support it.

Despite massive investments from big tech and rapidly growing demand for computing power, the entire supply chain appears to be under pressure: from semiconductor production to energy availability, as well as permits for new data center construction and shortages of electrical components and specialized labor.

In this context, even large-scale infrastructure projects are being slowed or suspended, as is the case with some initiatives in the United Kingdom, held back by high energy costs and regulatory complexities. The overall picture therefore indicates that the main constraint on the development of artificial intelligence is no longer financial or technological, but physical and infrastructural, with an increasing risk of delays or cancellations across the entire expansion chain.