Data centers expand globally

Data centers, which are essential infrastructures supporting the digital world, are expanding rapidly in France and globally. Their resource-intensive nature has raised concerns about their impact on land, water, and energy use, leading to local tensions over land use and electricity consumption.

Experts suggest new roles for data centers

Some experts suggest that the location and design of data centers could be reconsidered to transform them into contributors to the energy transition, producing heat and energy for local communities rather than being sources of conflict.

Growth amid economic challenges

The growth of data centers is occurring amid broader economic challenges, including rising prices, increasing debt, and sectoral crises. However, the artificial intelligence (AI) sector is experiencing rapid expansion.

AI drives data center spending

According to economist Jason Furman, construction spending in the United States has generally declined, except for data centers, which have seen a 64.5% increase in construction spending compared to the previous year. This growth is closely tied to the energy production required to power these facilities.

AI investment focuses on chips and operations

Expenditures related to AI development reveal a significant allocation pattern. For every 100 dollars invested in AI, only 7.5 dollars go toward infrastructure construction, while 65 dollars are allocated to the purchase of chips, memory, and network equipment. An additional 27.5 dollars are used for operations, primarily covering energy costs and cooling.

Hyperscalers face financial challenges

Companies managing the largest data centers, known as AI hyperscalers, are projected to invest 800 billion dollars in 2026, with this figure expected to rise to 1,100 billion dollars in 2027. However, the revenue generated from AI-related activities is currently far below the investment levels.

Profitability remains uncertain

Hyperscalers would need to generate 300 billion dollars annually from AI to amortize their investments, in addition to revenue from other activities. Presently, AI-related revenue stands at 70 billion dollars. To achieve a 30% profit margin, these companies would need to generate 636 billion dollars annually from AI, which is ten times their current revenue from this sector.

Investment surge may not be sustainable

The current situation is sustained by a surge in investment, but if this pace slows, it could lead to a potential AI investment bubble bursting.