Key Highlights

• AI is turning electricity availability into a strategic constraint for data-centre expansion.

• Global data-centre electricity consumption is projected to rise sharply through 2030.

• Grid connections, transformers, power electronics, storage and generation capacity are becoming central to AI infrastructure planning.

• Power infrastructure could become one of the widest non-technology beneficiaries of the AI Investment cycle.

Why has electricity become a central AI investment theme?

Artificial intelligence is creating an unusual connection between digital technology and physical energy infrastructure. Training and running advanced models require increasingly powerful data centres, and those facilities need reliable electricity around the clock. As AI workloads grow, the ability to secure power can become a constraint on where data centres are built and how quickly they can become operational. This is moving utilities, grid equipment, power generation and energy storage into the centre of the AI investment discussion.

How fast is data-centre electricity Demand expected to grow?

Updated energy-sector projections indicate that global data-centre electricity consumption could roughly double from about 485 terawatt-hours in 2025 to around 950 terawatt-hours by 2030. AI-focused facilities are expected to grow faster than the broader data-centre market. The figures illustrate why AI cannot be analysed solely through semiconductor or software demand. Electricity generation, transmission and distribution infrastructure must expand alongside computing capacity if the planned buildout is to be delivered.

Why are grids becoming a bottleneck?

Power generation may exist in a region without sufficient transmission capacity or a timely connection to a new data centre. Large AI facilities can require substantial loads, creating pressure on local substations, transmission lines and transformers. Grid planning can take years, while technology companies may want computing capacity much sooner. This timing mismatch can create bottlenecks and increase the value of sites with existing power availability and mature grid connections.

What role do transformers and power electronics play?

High-density computing requires sophisticated electrical infrastructure to deliver power efficiently and reliably. Transformers, switchgear, power-management systems and advanced power electronics are becoming increasingly important as rack-level power requirements rise. Some of these technologies already face long Manufacturing lead times. This creates a potential investment theme around specialist suppliers that can expand capacity while maintaining quality and reliability standards.

Could energy storage become part of the AI infrastructure solution?

AI data centres have unusual power characteristics because workloads can create large and rapid changes in electricity demand. Battery storage and other forms of energy management can help smooth those fluctuations and provide backup capacity. Longer term, storage could also allow data centres to participate more actively in electricity markets where regulations permit. The potential expansion of battery systems at data centres adds another layer to the AI infrastructure value chain.

What is the role of generation capacity?

New data centres may require additional generation from a combination of grid supply, renewable power, Natural Gas and other technologies depending on local conditions. The key requirement is dependable electricity at competitive cost. Companies developing generation capacity near major data-centre clusters could benefit from long-duration contracts, while utilities may need significant Capital Investment to expand networks and meet rising demand. The investment cycle therefore extends well beyond the technology sector.

What could slow the power-demand boom?

Permitting, local opposition, equipment shortages, grid constraints and higher financing costs could slow new projects. Energy affordability is another important issue because rapid demand growth can affect electricity prices for other consumers. Policymakers and regulators may therefore seek a balance between supporting data-centre investment and maintaining reliable, affordable power. These constraints could lead to a more gradual buildout in some regions even if global AI demand remains strong.

What should investors watch next?

The most important indicators include Utility capital-expenditure plans, grid-connection queues, transformer production, electricity prices, data-centre construction pipelines and storage deployments. Investors should also watch whether technology companies secure long-term power contracts and whether project timelines are being delayed by infrastructure limitations. The AI power theme is increasingly a structural infrastructure story: as computing demand rises, the companies able to provide reliable electricity and the equipment required to deliver it can become increasingly important to the global investment cycle.

The power theme also changes the way investors should think about AI infrastructure. A data centre cannot operate simply because processors and servers are available; it requires a complete ecosystem of generation, transmission, transformers, cooling, backup systems and grid management. This creates a longer-duration capital cycle that may continue even if individual technology products change. The pace of investment will vary by region, but the underlying requirement for reliable electricity is becoming a strategic component of the AI economy.