Today’s digital economy is placing unprecedented strain on the power grid. For years, efficiency gains have kept electricity use relatively stable despite rising demand, enabling net-zero ambitions.
But artificial intelligence, cloud and edge computing now threaten to reverse this trend. Indeed, forecasts predict a sharp rise in power demand through 2030.
The good news is these challenges are not insurmountable; by adopting proactive strategies such as alternative power sources, infrastructure planning and software, operators can secure capacity, build resilient facilities and scale sustainably.
The drivers of grid strain
AI is a leading contributor. A single AI search consumes about 10 times the electricity of a Google search, and forecasts suggest that AI and crypto mining could account for up to 10% of global electricity demand by 2030. Inference workloads, which outpace training clusters in energy needs, will accelerate consumption as models move into production.
Edge computing and IoT growth are also adding to consumption. The IEA suggests that “after globally consuming an estimated 460TWh in 2022, data centres’ total electricity consumption could reach more than 1 000TWh in 2026”, which is roughly equivalent to the electricity consumption of Japan.
Overall, customer grid connection queues in energy-constrained regions now often span three to five years, driven by rising demand and the clean energy transition. It is forecasted that the global power demand from data centres will increase 50% by 2027 and by as much as 165% by the end of the decade, while some major data centre hubs like Northern Virginia in the US have announced additional grid connection delays of up to seven years for large data centres amid a surge in requests.
Energy procurement strategies
Before the AI boom, data centres could tolerate longer utility lead times. Today’s urgent demand calls for alternatives such as natural gas turbines, hydrotreated vegetable oil-fuelled generators, wind, solar, fuel cells, battery energy storage systems (Bess) and small modular reactors (SMRs).
In their 2024 Midyear Global Outlook, BlackRock indicates both AI and low-carbon transition could spur historically large capital spending, as well as incremental investment to meet growing energy demand, with sources like solar farms, power grids, oil, and gas expected to generate investment of US$3.5-trillion (R61-trillion) per year this decade.
Today, two scenarios define the options:
Use case 1: Capacity exists but requires grid flexibility
Operators must provide energy storage to comply with demand response requests. Diesel generator sets remain common, but HVO fuels can save the day, offering a sustainable alternative that reduces carbon emissions by about 90%.
Behind the grid, Lithium-ion battery energy storage systems (Li-ion Bess) can provide enough energy storage to meet grid operator requirements for demand response and grid services, providing five key outcomes:
- Additional backup power
- Reduced diesel reliance
- Grid services participation
- Demand charge avoidance
- Greater use of renewables
Use case 2: No grid capacity exists
Here, operators need an alternative prime power. In general, natural gas turbines are the most used prime power alternative in the short term thanks to the technology’s maturity, cost, fuel flexibility, reliability, deployment speed and scalability.
SMRs, while still in the early stages of adoption due to regulatory hurdles, costs and public concern over the use of nuclear energy, are another option hailed as a future solution due to their promise of constant, carbon-free energy in a company’s footprint.
Finally, larger data centre providers could explore local microgrid providers to skip grid queues. These are companies with massive megawatt industrial parks that may already be grid-connected with microgrids as a backup or they may be permanently islanded, offering power-as-a-service.
Utilities and software: multipliers of efficiency
Beyond hardware, partnerships with utilities are vital. They offer important insights on grid capacity, carbon intensity, power quality and regulatory landscapes, all crucial for site selection, permitting and sustainable operations.
Furthermore, advanced technologies like digital twins and energy management software (EMS) enable better forecasting, efficient cooling, load balancing and overall optimisation. These tools also reduce waste, cut costs, and improve sustainability while maintaining performance.
Next steps for operators
To mitigate looming constraints, operators should:
- Reconsider site selection: Prioritise regions with strong grids or on-site generation opportunities while balancing latency and labour needs.
- Adopt alternative power sources: From natural gas turbines and HVO-fuelled generators to fuel cells, solar, wind and Bess.
- Collaborate early with utilities: Align on capacity, demand response and infrastructure sharing.
- Design for sustainability: Renewable integration improves efficiency, eases permitting, and strengthens reputation.