SUMMARY
Electricity is becoming one of the most important constraints on technological growth. Artificial intelligence, data centres, electric vehicles, semiconductor factories and advanced computing all require large and increasingly reliable supplies of power. The International Energy Agency projects global data-centre electricity consumption to roughly double from 2025 to 2030, while new research shows that access to power is becoming a critical factor in deciding where technology infrastructure can be built.
TECHNOLOGY — For decades, the technology industry was largely associated with processors, software, internet connections and data.
But another resource is becoming increasingly important: electricity.
The rapid expansion of artificial intelligence, cloud computing, data centres, electric vehicles, robotics and advanced manufacturing is creating a new challenge for technology companies. Building more powerful systems requires not only better chips and software, but also enormous amounts of reliable electricity.
This is changing the relationship between technology and energy.
Why Electricity Matters More Than Ever
Almost every modern digital service ultimately depends on electricity.
A smartphone requires power to operate. A telecommunications network needs electricity to transmit data. Cloud platforms depend on data centres filled with servers. AI models require powerful computing hardware that consumes electricity while training and operating.
As technology becomes more powerful, the amount of electricity required to operate the underlying infrastructure can increase significantly.
The International Energy Agency estimates that global data-centre electricity consumption was around 485 terawatt-hours in 2025 and could reach approximately 950 terawatt-hours by 2030.
Artificial Intelligence Is Changing the Equation
Artificial intelligence is one of the biggest reasons electricity has become strategically important to the technology industry.
AI systems are trained and operated primarily inside data centres. These facilities contain large numbers of processors, networking equipment, storage systems and cooling infrastructure.
The more sophisticated AI becomes, the more computing capacity companies need.
That creates a direct relationship between AI expansion and electricity demand.
The IEA says electricity consumption from AI-focused data centres grew by 50% in 2025, substantially faster than overall data-centre electricity consumption.
Data Centres Are Becoming Power-Hungry Facilities
A modern data centre is effectively a large industrial facility for computing.
Thousands of servers can operate continuously, generating significant heat that must be removed through cooling systems.
The electricity requirement therefore comes from more than the processors themselves.
Servers, networking equipment, cooling systems, power-management equipment and other infrastructure all contribute to the total load.
The IEA estimates that servers account for around 60% of electricity demand in modern data centres on average, although the figure varies by facility.
Power Is Becoming a Limitation on AI Expansion
Technology companies can design increasingly powerful AI systems, but those systems cannot operate without sufficient electricity.
This creates a potential bottleneck.
Building a new data centre can take only a few years, while expanding electricity generation and transmission infrastructure can require much longer planning and construction periods.
As a result, companies may find that the limiting factor is no longer simply whether they can purchase enough chips or build enough computing infrastructure.
They may also need to answer a more basic question: Where can they obtain enough reliable electricity?
Location Is Becoming an Energy Decision
In the past, technology companies often prioritised factors such as access to customers, fibre-optic networks, skilled workers and favourable business environments when choosing where to build infrastructure.
Electricity availability is becoming another major consideration.
A location with abundant, reliable and affordable electricity can become attractive for large data centres.
Conversely, a region with limited grid capacity may struggle to attract major AI infrastructure even if it has strong technology companies and excellent internet connectivity.
The U.S. Is Already Seeing the Pressure
The United States provides a clear example of how data-centre expansion is affecting electricity planning.
The U.S. Department of Energy's latest data-centre analysis estimates that data centres could account for 11.8% of total U.S. electricity consumption by 2030, with a range of possible outcomes from 9.5% to 15.3%.
That represents a significant change for an industry that historically occupied a much smaller share of electricity demand.
The Department of Energy has identified data-centre expansion, AI applications, manufacturing growth and electrification as important drivers of rising U.S. electricity demand.
Electricity and Semiconductor Manufacturing
AI is not the only technology increasing electricity requirements.
Semiconductor factories also require highly reliable power supplies.
Advanced chip manufacturing involves sophisticated machinery, controlled environments and continuous industrial processes.
An interruption in electricity supply can potentially disrupt production and create significant financial losses.
This means countries competing to attract semiconductor manufacturing need more than tax incentives and skilled workers. They also need reliable energy infrastructure.
Electric Vehicles Add Another Layer
The electrification of transportation is creating another source of electricity demand.
Electric vehicles shift a significant part of transportation energy consumption from petroleum products toward electricity.
As electric-vehicle adoption increases, charging infrastructure will require additional generation and grid capacity.
This creates an interesting connection between two major technology trends: AI increases electricity demand through computing, while electric vehicles increase demand through transportation.
Both are contributing to a broader shift toward an economy that depends more heavily on electricity.
Electricity Is Not Just About Quantity
Having enough electricity is only part of the problem.
Technology companies also need power that is reliable.
A data centre cannot simply shut down every time electricity supply becomes unstable.
AI systems and cloud services often operate continuously, meaning power interruptions can affect businesses, applications and critical digital services.
This makes grid reliability just as important as total electricity generation.
Why Transmission Networks Matter
Generating electricity does not automatically solve the problem.
The electricity must also reach the location where it is needed.
That requires transmission lines, substations, transformers and other grid infrastructure.
In some regions, electricity may be available somewhere within the country while the local grid lacks the capacity to deliver enough power to a new data centre.
This creates another potential bottleneck for technology expansion.
Renewable Energy and Technology
Technology companies are increasingly interested in renewable electricity because of both environmental commitments and the need to secure additional power.
Solar and wind projects can provide large quantities of electricity, although their output varies according to weather and time of day.
This means renewable generation may need to be combined with storage, grid connections, flexible generation or other technologies to provide reliable electricity for large computing facilities.
The IEA expects renewables to meet nearly half of the additional electricity demand from data centres through 2030 in its base case.
Natural Gas Is Also Part of the Picture
Natural gas is expected to remain important in some markets because gas-fired power plants can provide electricity when demand is high and renewable generation is unavailable.
The IEA expects renewables, natural gas and eventually nuclear power to play important roles in meeting rising data-centre electricity demand.
This creates a complicated energy debate for technology companies.
They need electricity that is affordable and dependable while also facing pressure to reduce emissions.
Nuclear Power Is Returning to the Technology Conversation
The growth of AI and data centres has also renewed interest in nuclear energy.
Nuclear power can provide large quantities of electricity continuously without the direct carbon emissions associated with fossil-fuel combustion during generation.
For technology companies that require dependable power around the clock, that characteristic can be attractive.
However, nuclear projects can take years to develop and involve major regulatory, financial and engineering challenges.
Power Density Is Becoming a New Technology Challenge
AI hardware is becoming increasingly powerful, but that creates another problem: power density.
More computing capability is being packed into relatively small spaces, requiring more electricity and more sophisticated cooling.
The IEA reports that the power density of AI servers increased dramatically between 2020 and 2025 and expects further increases by 2027.
This means future data centres may require not only more electricity overall but also much more electricity delivered to individual racks and computing systems.
Cooling Is Part of the Electricity Problem
Computing equipment generates heat.
As processors become more powerful, removing that heat becomes increasingly important.
Cooling systems therefore consume additional electricity and can become a significant component of a data centre's overall energy requirements.
Technologies such as liquid cooling are being developed to handle higher-density computing more efficiently, but they also introduce new engineering requirements.
Electricity Could Become a Competitive Advantage
This creates a major strategic shift.
Countries with abundant electricity, strong grids and competitive energy prices could become attractive locations for AI infrastructure.
Countries with limited electricity generation or weak transmission networks may find it harder to compete for large data-centre investments.
In other words, electricity infrastructure could increasingly influence which countries become major hubs for artificial intelligence and advanced computing.
The AI Race Is Also an Energy Race
It is tempting to describe the AI competition entirely in terms of algorithms, chips and data.
But the infrastructure underneath AI is equally important.
A company can have access to advanced processors and world-class researchers, but without sufficient electricity it cannot operate large-scale computing infrastructure.
The U.S. Department of Energy explicitly links AI leadership with the need to expand energy infrastructure and computing capacity.
This suggests that future competition in AI will increasingly involve energy policy as well as technology policy.
Efficiency Could Change the Outlook
There is an important counterargument to the idea that technology will simply consume ever-increasing amounts of electricity.
Computing hardware and software are becoming more efficient.
The IEA notes that energy use per individual AI task has fallen dramatically as hardware and software improve. At the same time, more energy-intensive applications such as AI video generation, reasoning and agentic systems are emerging.
This creates a race between efficiency improvements and increased usage.
If efficiency improves faster than demand grows, the electricity impact could be reduced. If AI adoption expands faster, total electricity consumption could continue rising rapidly.
The Grid Must Become More Flexible
Future electricity systems will need to manage increasingly complex demand.
AI data centres can create large and concentrated electricity loads, while renewable generation can fluctuate depending on weather conditions.
Energy storage, demand management, improved transmission networks and advanced grid technologies could therefore become increasingly important.
The challenge is not simply generating more electricity. It is matching electricity supply with rapidly changing demand at the right place and time.
Why This Matters to Consumers
The consequences could eventually reach ordinary electricity customers.
If large technology facilities require major grid upgrades, questions arise over who should pay for those investments.
Governments and regulators must balance the economic benefits of new data centres with the potential costs to electricity systems and existing customers.
The U.S. Department of Energy has specifically highlighted the need to expand infrastructure while protecting affordability and reliability for households and businesses.
A New Definition of Technology Infrastructure
Technology infrastructure used to mean servers, fibre-optic cables, software platforms and computer chips.
Increasingly, the definition is expanding.
It now includes power plants, transmission lines, substations, transformers, batteries, cooling systems and other energy infrastructure required to keep digital systems operating.
This means the technology industry and the energy industry are becoming more closely connected than ever.
The Bigger Picture
The most important change is not that technology suddenly uses electricity.
Technology has always depended on power.
The difference is scale.
AI, cloud computing, semiconductor manufacturing, electric transportation and advanced automation are expanding simultaneously. Together, they are creating a future in which electricity availability can determine how quickly new technologies are deployed.
The IEA expects global data-centre electricity consumption to roughly double from 2025 to 2030, reaching about 950 TWh in its central outlook.
Conclusion
Electricity is becoming a strategic resource for technology because modern computing is becoming increasingly energy-intensive.
Artificial intelligence is accelerating demand for data centres, while semiconductor factories, electric vehicles, robotics and other technologies are adding further pressure to electricity systems.
This does not mean the world is necessarily running out of electricity. The bigger challenge is building generation, transmission and grid infrastructure quickly enough, and in the right locations, to support rapidly growing technology demand.
The countries that can provide abundant, reliable and affordable electricity may gain an important advantage in the next phase of technological development.
The future technology race will therefore not be fought only over chips, algorithms and data. Increasingly, it will also be fought over something much more fundamental: who has enough power to run the machines.
