Data centres are the new energy factories

This article is part of In conversation about sustainable finance & emission reduction systems, a new series by Diego Balverde. You're reading volume twelve of the Energy Shocks series. Here is volume eleven

Part of Diego Balverde's upcoming book on how wars, gas, electricity and infrastructure are redrawing the global economy.

Prologue

Data centres are no longer digital buildings. They are energy factories. They consume electricity at industrial scale, compete for grid capacity, reshape local infrastructure, influence water demand, alter real estate markets and force governments to decide who gets power first.

The global economy is entering a phase in which artificial intelligence, cloud computing, financial markets, defence systems, logistics, health data and industrial automation all depend on facilities that cannot function without massive, stable and continuous electricity. This means the digital economy is no longer weightless. It has a physical body made of power lines, substations, cooling systems, backup generation, land, water, chips, batteries and capital.

The next energy shock will not only hit factories and households. It will hit the servers that increasingly run the economy.

The digital economy is becoming an electricity economy

For years, the digital economy was described as immaterial. Software, platforms, cloud services and artificial intelligence seemed to float above the physical world. That idea is now obsolete. Every AI model trained, every cloud service delivered, every financial transaction processed, every logistics system optimised and every industrial platform monitored depends on electricity. The digital economy does not live in the air. It lives in data centres connected to grids.

Global electricity demand already exceeds 30,000 TWh per year and continues to grow above 4% annually. Data centres alone are moving toward consumption above 1,000 TWh annually, a level comparable to the electricity demand of a major industrial economy.

That figure changes the debate. Data centres are not marginal consumers. They are becoming industrial loads. They do not behave like ordinary commercial buildings. They require firm power, redundancy, cooling, backup systems, high uptime and rapid connection. Their electricity demand is concentrated, continuous and difficult to interrupt.

This creates a new conflict inside the energy system. A region may want data centres because they attract investment, technology companies, digital infrastructure and political prestige. But the same region may also need electricity for households, industry, ports, rail, electric vehicles, hydrogen projects, battery factories and public services. When the grid is limited, not everyone can grow at the same time. The question becomes political and financial: who receives the next available megawatt.

The real issue is not whether data centres are good or bad. The issue is whether energy systems are prepared for them. If data centres arrive faster than grids, storage and clean generation, they intensify bottlenecks. If they are integrated with storage, efficiency, demand response, heat reuse, distributed generation and finance, they can become part of the infrastructure solution. The difference is design.

The data shows a new industrial load

The scale of the data centre boom is difficult to overstate. AI workloads require far more electricity than traditional computing. Training large models, operating inference at scale, cooling high-density server racks and maintaining redundancy all increase demand. A single hyperscale data centre can require hundreds of megawatts. A cluster of them can demand power at the scale of an industrial zone.

This is why grid operators are beginning to treat data centres less like office buildings and more like steel plants, refineries or chemical complexes.

The pressure does not end at electricity. Cooling systems can raise water and energy demand. Backup generators increase fuel exposure. Battery systems require critical minerals and supply chains. Transmission upgrades require transformers, cables, land, permits and financing. Substations require time. In some regions, grid connection queues already stretch years. That means the limiting factor for data growth may not be chips or software. It may be electricity delivery.

This changes the economics of technology. A technology company may have capital, customers and computing demand, but without firm power its growth becomes constrained. A region may have land and political ambition, but without grid capacity it cannot host the next wave of digital infrastructure. A utility may welcome large demand, but if it cannot expand transmission and balancing capacity, the new load can raise prices for everyone else. Data centres therefore become a test of whether the energy system can absorb digital growth without destabilising the wider economy.

The financial implications are enormous. Electricity price volatility can directly affect the operating cost of data centres. Grid congestion can delay projects. Cooling demand can intensify during heat waves precisely when the grid is already under stress. Backup fuel exposure can rise during supply disruptions. If capital markets begin pricing energy risk into digital infrastructure, data centre valuation will increasingly depend on energy strategy, not only on computing demand.

When servers compete with factories, the economy must choose

The rise of data centres creates a new form of competition inside the electricity system. It is not only competition between companies. It is competition between economic models. Should a constrained grid prioritise a data centre, a battery factory, a port electrification project, a hydrogen plant, a residential district or an industrial cluster. This is no longer a technical queue. It is an economic choice.

If data centres absorb large volumes of firm power without contributing flexibility, they can intensify congestion and raise costs. If they are built in regions where the grid is already stressed, households and industry may face higher prices or delayed connections. If they rely heavily on backup diesel during stress events, they can weaken emissions goals. If they do not integrate storage or demand flexibility, they become large rigid loads in a system that increasingly needs flexibility.

But the opposite is also true. If designed correctly, data centres can accelerate energy infrastructure. They can anchor new renewable projects, finance grid upgrades, support storage deployment, reuse heat, provide demand response in limited windows and create long-term offtake structures that make clean energy projects bankable. The question is not whether data centres consume energy. The question is whether they are integrated into the system as passive consumers or active infrastructure partners.

This is where the politics becomes difficult. AI, cloud infrastructure and digital services are now strategic. Governments want them for competitiveness and security. But citizens want affordable electricity. Industries want stable power. Climate targets require lower emissions. Banks want bankable infrastructure. Utilities want predictable demand. If these objectives are not aligned, data centres become a new source of tension.

The energy shock of the digital economy is therefore not only a demand shock. It is an allocation shock. It forces societies to decide how electricity should be used, priced, financed and governed.

The new model is compute, storage, grid and finance

The solution is not to stop data centres. The solution is to integrate them into energy system design. That means every major data centre strategy should include grid impact analysis, distributed generation, storage, cooling efficiency, heat reuse, MRV, demand flexibility, local workforce training and financial structures that reduce pressure on the wider system.

BalGreen's architecture is relevant in this context because it treats energy demand as something to be structured, not simply supplied. Modular panelisation using mathematical optimisation can accelerate distributed generation around industrial and digital sites without revealing the full method. Storage can reduce peak demand exposure and support grid stability. Training programmes can create local technical capacity for installation, maintenance and monitoring. MRV can verify efficiency gains, emissions reductions and operational performance. Ports, logistics hubs and industrial parks can be linked with data infrastructure to create integrated energy zones rather than isolated high-demand islands.

NatureAlpha can support environmental intelligence and exposure mapping, identifying where climate risk, grid stress, water pressure and asset vulnerability intersect. StoneX can support energy price-risk management, hedging and market execution where electricity and fuel volatility affect operating costs. BlackRock and Standard Chartered can support large-scale capital structuring when digital-energy infrastructure becomes standardised and bankable. Gold Standard can strengthen credibility around verified emissions reductions and climate-linked monetisation.

The value is in turning data centres from passive load into system assets. A data centre that only consumes power creates pressure. A data centre that finances generation, integrates storage, manages peak exposure, verifies emissions and supports grid resilience creates value. The difference is not the server. It is the architecture around the server.

This also creates a new investment thesis. The next premium will belong to locations that can offer clean, stable, financeable electricity with low grid risk. Regions that can combine land, power, storage, water management, digital infrastructure and financing will attract the next wave of AI and cloud investment.

Regions that cannot will lose projects even if they offer tax incentives.

Analysis

If data centres are moving toward more than 1,000 TWh of annual electricity demand, why are they still discussed as technology assets rather than energy assets? If a single data centre cluster can require power at industrial scale, should it compete equally with factories, ports and households for grid access? If AI needs continuous electricity, who controls the real infrastructure of artificial intelligence: chipmakers, cloud companies or grid operators?

If data centres raise local demand faster than grids expand, are they accelerating digital growth or importing energy fragility? If a region offers tax incentives but lacks firm power, is it really competitive? If backup diesel protects uptime but increases emissions exposure, should data centres be required to finance cleaner resilience?

If storage can reduce peak pressure, why is it not mandatory for large digital loads? If waste heat can be reused, why is it still wasted in many systems? If MRV can verify lower emissions and efficiency, why is digital infrastructure not more directly connected to climate finance?

If data centres become strategic infrastructure, who should govern their energy footprint: markets, regulators, utilities or national security agencies? And if the digital economy depends on electricity, why do we still talk about AI as if it were immaterial?

Conclusion

My conclusion is clear. Data centres are the new energy factories because the digital economy now depends on massive, continuous and financeable electricity. Artificial intelligence, cloud systems, logistics platforms, financial markets and defence infrastructure all require power before they create value. The future of digital leadership will not be decided only by algorithms, chips or capital. It will be decided by electricity access, grid capacity, storage, cooling efficiency and energy finance.

The next winners will be the regions and companies that understand that computers are energy. The next losers will be those who build digital ambition without power architecture. In the new economy, the server is not weightless. It is connected to the grid. And the grid will decide who scales.

illuminem Voices is a democratic space presenting the thoughts and opinions of leading Sustainability & Energy Thought Leaders, their opinions do not necessarily represent those of illuminem.

Track the real-world impact behind the sustainability headlines. illuminem's Data Hub™ offers transparent performance data and climate targets of companies driving the transition.