Key takeaways
For more than two decades, digital transformation was seen as a largely immaterial process. The vocabulary itself (cloud, virtualization, dematerialization) fed the idea of an economy freed from physical constraints. That view no longer holds, and this is becoming more and more obvious.
The rise of artificial intelligence marks a tipping point. It reveals what digital technology has always been, without ever really wanting to admit it: an industrial activity, dependent on energy, natural resources, physical locations and, increasingly, the geopolitical balance of power.
The global rush for data centers is therefore not merely a technical phenomenon. It is strategically revealing. It exposes the limits of a digital transformation designed âin a vacuum,â cut off from its material foundations.
When computing becomes the constraint
Since 2022, tech has been changing the very nature of the digital constraint. For a long time, the limiting factor was software: the conversation was about integration capacity, system maturity and user adoption. Today, the constraint is physical and material. Computing consumes electricity, and at scale, that consumption becomes a defining factor.
According to the International Energy Agency (IEA), data centers accounted for about 460 TWh of global electricity consumption in 2022, or nearly 1% of global demand. This figure, which had been relatively stable for several years, could more than double by 2030, driven by the combined growth of generative AI, cloud computing and massive data storage.
This is not just a matter of volume. Geographic concentration is key. Hyperscale data centers, whose installed capacity now reaches 50 to 200 MW per site (the equivalent of a mid-sized city), create significant local constraints: grid connection, water availability, community acceptance, permitting timelines.
Digital performance now depends on the ability to secure finite physical resources.
A data center is not an IT component: it is a factory!
A modern data center is a heavy industrial asset. It requires land, concrete, steel, copper, electrical equipment, batteries, generators and complex cooling systems. It depends on global supply chains, which are often highly concentrated.
Water is becoming a critical factor. Depending on the cooling technology, a large data center can consume several million cubic meters of water per year. Google, one of the few companies that publish this data, reported total water consumption of 21 billion liters in 2022, a significant increase, largely linked to its operations and its data centers.
Electricity is the lifeblood of the sector. In some regions, such as Ireland, Northern Virginia and Singapore, authorities have begun to restrict or place conditions on new projects, not for technological reasons, but because power grids can no longer keep up. The US Department of Energy is now warning about the growing pressure AI is putting on energy planning.
In effect, digital transformation is becoming a matter of energy policy.
An international race for computing power
The rush for data centers reveals a quiet geopolitical realignment.
- The United States accounts for nearly 40% of global data center capacity, driven by the vertical integration of Big Tech hyperscalers (Amazon, Microsoft, Google), privileged access to capital and massive investment capacity. These companies secure their own energy, sites and supply chains directly.
- China, for its part, takes a different approach: from land-use planning to state control of land and energy, not to mention the accelerated development of domestic computing capacity to reduce its technological dependence.
- Europe, despite strong ambitions on digital sovereignty, remains structurally fragmented and lagging behind. Energy, regulatory and land constraints slow deployment, while most critical computing capacity remains controlled by non-European players.
Other regions, notably the Middle East and Southeast Asia, are investing heavily to capture this new strategic value. Computing is becoming a geopolitical asset, just like energy or telecommunications.
Sovereignty: a matter for states⌠and for companies
Digital sovereignty is often framed as a matter for governments. But for companies, it is becoming an operational issue. It is not about bringing everything in-house or rejecting the cloud. It is about answering questions that can no longer be avoided:
- Where is critical data stored?
- How dependent are we really on a single supplier, country or energy region?
- What happens in the event of an energy, regulatory or geopolitical shock?
- What is the emergency and backup plan in case of failure?
In a world of constraints, business continuity, compliance and control of the data value chain are becoming top strategic priorities.
Twin Transformation: finally ending âdisconnectedâ management
Twin Transformation (bringing digital and sustainability together) is often presented as a virtuous convergence. In reality, it requires a deeper paradigm shift. We have to accept that digital now grows within the limits of the physical world. The rush for data centers highlights three mistakes inherited from past digital transformations.
- The first is confusing strategy with the accumulation of technology. Deploying AI without thinking about energy simply shifts the constraint rather than solving it.
- The second is underestimating dependencies. Digital technology relies on a complex value chain (energy, rare components and minerals, networks, skills and the right leadership) whose fragility is becoming a major strategic factor. The IMF notes that the technology transition is intensifying competition for rare minerals, with high geographic concentration and a heightened risk of crises. The Venezuela hold-up in early 2026 is just one example of the major risks ahead.
- The third is treating sustainability as a compliance and paperwork exercise, when it is becoming a design parameter: where infrastructure is located, architecture choices, sober usage, AI governance.
What this demands of leaders
- For executives, the challenge is not to add a layer of responsible rhetoric, also known as âgreenwashingâ. It is to revisit certain fundamental decisions. First, map the companyâs real dependence on digital and energy infrastructure, not as a CSR indicator, but as a business risk.
- Then, define what is critical and sovereign for the organization: sensitive data, key models, operational continuity. And follow through on what that means for architecture, location and partnerships.
- Finally, make a clear-eyed trade-off between power and sobriety. Tech, including AI, is a strategic lever, not an end in itself. Its value is measured not by the amount of computing power it uses, but by the real advantage it creates.
The rush for data centers is a moment of truth. It forces us to see digital transformation for what it has become: industrial, energy-intensive, territorial and geopolitical. And that is exactly why it is an âexpertâ topic: infrastructure reveals what we were trying not to see, a bit like the elephant in the room.
Twin Transformation is not âdoing sustainability on top of digital.â It means accepting that digital must now be designed for a finite world. Keep in mind that a tech strategy that does not know where it lives is not a strategy: it is a mere directive. In a world of constraints, abstraction becomes an unmanaged risk.
Sources
- International Energy Agency (IEA): Electricity 2023, Data centres and AI
- U.S. Department of Energy: AI, Data Centers and Grid Reliability
- Google: Environmental Report 2023 (water & energy usage)
- International Monetary Fund (IMF): Critical Minerals and Geopolitical Risk
- European Commission: Energy Efficiency Directive & Data Centres
- European Commission / EU framework (reporting, energy efficiency, etc.: regulatory context for data centers)
- Reuters: coverage of data center expansion and energy constraints (EU, France, US)

Malika Ait El Mouden explores how technologies, organizations and living systems can transform together. A digital transformation specialist with 18 years of experience in Europe, the Middle East and Southeast Asia, she has led international eCommerce and omnichannel programs. She has founded two companies, including a start-up that contributed to the development of circular solutions.
A board member of La French Tech and of the APM (Association pour le Progrès du Management), she also teaches in higher education and is following a program with MIT on the challenges of data and AI. She popularizes and spreads the concept of Twin Transformation.
Malika is the co-founder, with Julie Bonazzi, of Twin Transformation.fr



