Editor’s note: FM Perspectives are industry op-eds. The views expressed are the authors’ and do not necessarily reflect those of Facilities Management Advisor.
As artificial intelligence (AI) accelerates the construction of data centers, utilities are rapidly investing in energy infrastructure, including new power generation, transmission lines, and grid upgrades, to keep pace with the surge in energy demand. These investments ultimately raise costs for all customers. In regions experiencing concentrated data center growth, analysts project that residential electricity bills could rise by $18 per month. Meanwhile, industrial and commercial facilities face higher rates as utilities seek to recover infrastructure costs.
A single hyperscale AI data center can require 100 to 500 megawatts (MW) of electricity. This is roughly equivalent to the power demand of 80,000 to 400,000 homes. In 2024, U.S. data centers consumed 183 terawatt-hours (TWh) of electricity—more than 4% of total U.S. electricity use. Projections indicate that by 2028, this figure could rise 133% to 426 TWh. The International Energy Agency suggests that data centers may account for nearly half of all U.S. electricity demand growth between now and 2030.
At the facility level, the electricity demand is even more striking. With requirements ranging from 100 to 500 megawatts, a hyperscale data center often emerges as one of the largest new loads on regional power grids.
The need for efficiency has become urgent and unavoidable. As electricity is already one of the largest operating expenses for many industrial facilities, the stakes have never been higher for protecting ratepayers. However, focusing solely on prices falls short of addressing the immediate and escalating structural demand challenge.
Supply-side pledges address the symptoms, not the core issue. If efficiency is not central to the strategy, AI’s rapidly growing energy and water demands will keep outpacing infrastructure growth. This will drive up costs and strain both grid and water systems, which is an issue that cannot be fully remedied with new generation alone.
Rate protections may buy a brief reprieve, but they do nothing to halt the rapidly worsening trajectory of consumption or relieve the mounting strain on critical infrastructure. Delay risks compounding the crisis.
The numbers are stark.
Meanwhile, new power plants and transmission expansion take years to permit, finance, and build, but efficiency upgrades can be deployed in months. That timing gap is critical: While large-scale infrastructure projects move on multi-year timelines, demand from AI workloads scales based on a different pace.
Without systemic efficiency improvements now, supply will continue to lag behind surging demand, and ratepayers will face relentless volatility. Acting immediately is essential to avoid future crises.
Electricity, however, is only part of the challenge. Water systems—often managed locally and already under strain in many regions—are facing parallel pressures as data center demand accelerates.
A typical data center uses 300,000 gallons of water per day, while large facilities can consume up to 5 million gallons daily, equivalent to the water use of a town of 50,000 people. Across the U.S., data centers consumed an estimated 17 billion gallons of water for cooling in 2023. By 2028, that figure is expected to double or even quadruple.
Water scarcity, unlike electricity, cannot be managed solely with pricing. Local water supplies face physical limits; once resources are depleted, no financial tools can restore them. In drought-prone regions in particular, rising data center demand is likely to intensify competition among industrial and commercial users, municipal systems, and residential communities.
This underscores the need to address the energy-water nexus—the interconnected relationship between water and energy systems. Energy is required to pump, treat, and distribute water, while water is essential for cooling power plants and data centers. Efficiency gains compound across both systems. Less water use means less energy needed to treat and pump it. Lower energy demand means less heat to reject, and less water required for cooling. Treating these systems separately risks missing the compounding benefits of coordinated efficiency.
The implications demand immediate attention. Cooling systems account for the largest share of a data center’s water and energy use, making them the most urgent strategic leverage point for efficiency improvement. Advanced cooling retrofits, water reuse, heat recovery, and real-time monitoring can all deliver rapid and substantial resource reductions—without delaying AI growth. However, urgent action is needed beyond the data center itself.
With mounting pressure on energy and water supplies, rapidly improving efficiency across surrounding commercial buildings and industrial facilities within the same utility service areas is essential. Freeing up grid and water capacity can be immediately redirected to support AI infrastructure, intensifying the need to act before new power plants become unavoidable. A coordinated portfolio of efficiency investments—within data centers and across the broader service area—offers one of the fastest and most cost-effective routes to relieve capacity constraints, reduce capital expenditures, accelerate implementation, and lower overall system risk.
Efficiency is not optional and is becoming a central component of managing the infrastructure demands created by AI growth. While measures to protect ratepayers may help moderate short-term impacts, long-term stability will depend on sustained investment in systemic efficiency improvement. Without these efforts, pressure on electricity costs, grid capacity, and water resources will only intensify. As AI infrastructure continues to expand, the key challenge will be ensuring we manage AI growth wisely by supporting energy and water systems capable of reliably and efficiently sustaining it, without overwhelming the systems that sustain it.
Ya-Chi Tsao, Ph.D., P.E., is a licensed senior efficiency engineer at Ecosave specializing in energy systems for water and wastewater infrastructure. She received her doctorate in electrical and systems engineering from the University of Pennsylvania.