The digital world is growing at an extraordinary pace. Artificial intelligence, cloud computing, streaming, online services and increasingly sophisticated applications are transforming the way we work, communicate and consume information.

Behind all of this digital activity, however, there is something very physical: electricity.

Every search, image generated by artificial intelligence, video streamed, cloud application or business transaction requires computing power. And that computing power is concentrated in increasingly large facilities: data centers, or data processing centers (DPCs).

The question is no longer simply how much energy these facilities consume, but where that energy will come from and how we can make their growth compatible with a lower-carbon electricity system.

The answer is likely to involve a combination of renewable energy, storage, stronger electricity grids, nuclear power and much greater efficiency.

Artificial intelligence changes the scale of the challenge

The rapid expansion of artificial intelligence is accelerating the growth of data centers.

AI applications require large amounts of computing power, particularly during the training and operation of advanced models. As these technologies become integrated into search engines, office applications, industrial processes, autonomous systems and everyday digital services, electricity demand is expected to increase significantly.

According to the International Energy Agency (IEA), global electricity consumption by data centers was around 485 TWh in 2025.

The IEA expects this figure to approach 950 TWh by 2030, meaning that consumption could almost double in just five years.

By 2030, data centers could therefore account for around 3% of global electricity demand.

These figures are global averages, but the impact will not be evenly distributed. Data centers are concentrated in particular regions and increasingly compete for access to electricity, land, transmission capacity and cooling resources.

A global energy challenge

The growth of data centers is taking place at the same time as the world’s electricity system is undergoing a profound transformation.

Solar and wind power are expanding rapidly, battery storage is becoming increasingly important, electricity networks are being modernised and many countries are attempting to reduce their dependence on fossil fuels.

This creates both a challenge and an opportunity.

On the one hand, data centers represent a new and rapidly growing source of electricity demand.

On the other, their large and predictable consumption can become an important driver for new renewable projects and investments in electricity infrastructure.

The two developments are therefore increasingly connected.

Renewable energy will play a central role

Renewables are already an important part of the electricity supply serving data centers.

The IEA estimates that renewable sources currently cover around 27% of the electricity physically consumed by data centers worldwide.

Looking ahead, renewables could provide roughly half of the additional electricity required by data centers through 2030.

Solar photovoltaic power is particularly well positioned because of its rapidly falling costs and short construction times. Wind power can complement solar production by generating electricity during different periods of the day and year.

Hydropower can provide an additional source of flexibility in countries where it is available.

But there is an important limitation.

A data center cannot simply consume electricity when the sun is shining or the wind is blowing. Its servers operate continuously.

This brings us to one of the key concepts shaping the next stage of the energy transition.

From buying green electricity to 24/7 clean energy

Many technology companies already sign Power Purchase Agreements (PPAs) with renewable-energy producers.

These agreements can support the development of new solar and wind farms and allow companies to match their annual electricity consumption with an equivalent amount of renewable generation.

But annual matching is not the same as having clean electricity available every hour.

A data center may sign a contract for renewable electricity while still relying on the electricity grid at night or during periods of low wind and solar production.

This is why the concept of 24/7 Carbon-Free Energy is gaining importance.

The objective is much more demanding: ensuring that electricity consumption is matched with carbon-free electricity on an hourly basis.

Achieving this will require a broader energy mix.

Solar and wind will remain essential, but they will increasingly need to work alongside batteries, hydroelectric power, nuclear energy, demand flexibility and stronger electricity grids.

Batteries become part of the equation

Energy storage is becoming one of the most important pieces of this puzzle.

Battery installations grew by around 40% in 2025, with approximately 110 GW of new battery capacity installed globally.

For data centers, batteries can perform several functions.

They can provide backup power, help manage peaks in demand, absorb excess renewable generation and reduce pressure on the electricity grid.

In the future, large data centers could therefore operate not simply as electricity consumers, but as increasingly flexible participants in the power system.

This is an important change in perspective.

The electricity grid: perhaps the real bottleneck

Generating renewable electricity is only part of the challenge.

That electricity must also reach the places where data centers are being built.

Transmission networks in many regions are already under pressure. New generation projects can sometimes be constructed faster than the grid infrastructure required to connect them.

This creates a potential bottleneck.

A data center may have access to a large amount of renewable electricity on an annual basis, but if the local grid cannot deliver sufficient power at the required time, the project may face delays or additional costs.

For this reason, the expansion of data centers is also becoming a major driver of investment in:

• transmission lines;
• substations;
• transformers;
• distribution networks;
• energy storage;
• grid management systems.

The energy transition is therefore not only about building more renewable generation. It is also about building the infrastructure that connects it all together.

Location will become increasingly important

Where a data center is built can have a major impact on its energy footprint.

Access to abundant renewable electricity, reliable grids, suitable land and efficient cooling systems can make some locations considerably more attractive than others.

This could encourage the development of new data centers close to areas with abundant renewable resources.

In Europe, for example, the IEA expects renewables and nuclear power together to provide around 85% of the additional electricity required by data centers through 2030.

This illustrates an important principle: the future of digital infrastructure will increasingly be linked to the geography of energy.

The water challenge

Electricity is not the only resource involved.

Large data centers also require sophisticated cooling systems, and some cooling technologies consume significant amounts of water.

As facilities become larger and computing densities increase, cooling becomes an increasingly important part of their environmental footprint.

This is encouraging innovation in areas such as:

• liquid cooling;
• closed-loop cooling systems;
• more efficient heat exchangers;
• waste-heat recovery;
• alternative cooling technologies.

In regions affected by water scarcity, the choice of cooling technology could become an important factor when deciding where new data centers should be built.

And what about nuclear energy?

Renewables will be central to the future electricity supply of data centers, but they will not necessarily be the only solution.

Nuclear power is attracting renewed interest because it can provide large quantities of low-carbon electricity continuously, regardless of weather conditions.

For electricity-intensive facilities operating 24 hours a day, this characteristic is particularly relevant.

Some technology companies are therefore exploring long-term agreements involving existing nuclear plants, while others are examining the potential role of Small Modular Reactors (SMRs).

However, nuclear power also faces challenges, including construction times, financing, regulation, waste management and public acceptance.

The emerging energy system is therefore unlikely to depend on a single technology.

Artificial intelligence: part of the problem, but potentially part of the solution

There is an important paradox at the heart of the AI-energy relationship.

Artificial intelligence is driving additional electricity demand, but AI can also help reduce energy consumption elsewhere.

AI-based systems can optimise electricity networks, improve renewable-energy forecasting, detect equipment problems, optimise industrial processes and manage buildings more efficiently.

In other words, AI can simultaneously be a new source of energy demand and a tool for improving energy efficiency.

The challenge will be ensuring that the efficiency gains generated by AI are not outweighed by the additional energy required to run it.

A huge opportunity for renewable energy

The rapid expansion of data centers could become an important catalyst for the energy transition.

Technology companies require large amounts of reliable electricity and increasingly want to reduce the carbon intensity associated with their operations.

This creates strong incentives to invest in new renewable generation, storage and grid infrastructure.

In this sense, the digital revolution and the energy transition are beginning to converge.

The data center of the future may be directly connected to a solar farm, supported by wind power, backed by batteries, supplied through an upgraded electricity grid and complemented by firm low-carbon generation.

The objective is not necessarily to choose one technology over another, but to combine them intelligently.

The risk of increasing fossil-fuel dependence

There is, however, another possible outcome.

If electricity demand from data centers grows faster than clean-energy generation and grid infrastructure, fossil fuels could continue to play an important role in meeting the additional demand.

Natural gas, and in some regions coal, can provide dispatchable electricity when renewable generation is insufficient.

This would allow digital infrastructure to continue expanding, but it could also make decarbonisation more difficult.

The timing therefore matters.

It is not enough to build more data centers and add renewable capacity eventually. The electricity infrastructure needs to develop at roughly the same pace as digital demand.

What will the data center of the future look like?

The data center of the future will probably look very different from the traditional model.

Rather than being simply a large building connected to the electricity grid, it could become part of a much more integrated energy ecosystem.

We may see combinations such as:

Solar + wind + batteries + grid + nuclear + advanced cooling + energy efficiency.

The exact combination will depend on the location, available resources, regulations and economic conditions.

But the underlying principle will be the same: producing and delivering large amounts of reliable electricity while reducing environmental impact.

Efficiency will be as important as generation

There is another aspect that should not be overlooked.

The cleanest electricity is, in many cases, the electricity that does not need to be generated in the first place.

Improving the efficiency of processors, servers, cooling systems and data-center architecture will therefore be essential.

More efficient AI models, specialised chips, better software and improved thermal management can all reduce the amount of electricity required for a given computing task.

This means that the future of data centers will not depend solely on producing more electricity.

It will also depend on using electricity more intelligently.

The beginning of a new relationship between technology and energy

For decades, digitalisation was associated with the idea that information was becoming increasingly independent from the physical world.

Today we are discovering the opposite.

Behind every cloud service, AI model, digital image and online transaction there is a physical infrastructure that requires electricity, cooling systems, buildings, networks and resources.

The spectacular growth of artificial intelligence is making this reality impossible to ignore.

The challenge ahead is therefore not to stop digitalisation, but to make it increasingly compatible with a cleaner and more efficient energy system.

Renewables will play a fundamental role. Batteries will provide flexibility. Electricity grids will need to expand and become smarter. Nuclear power may provide additional firm low-carbon generation in some regions. And efficiency will become increasingly important.

The future will not be powered by a single technology.

It will be powered by integration.

The data center of tomorrow could become one of the most important meeting points between the digital revolution and the energy transition.

And the key question is no longer simply how much computing power we can build.

It is how intelligently we can connect that computing power to the energy system of the future.

---

Main sources

• International Energy Agency (IEA)
• IEA — Energy and AI
• IEA — Energy Demand from AI and Data Centres
• International Renewable Energy Agency (IRENA)