The rapid expansion of artificial intelligence infrastructure is forcing the power sector to reconsider how electricity networks support large-scale computing demand. Data centers, once viewed primarily as digital infrastructure assets, are becoming major energy consumers that influence utility planning, transmission investment and consumer electricity costs. Researchers at Lawrence Berkeley National Laboratory estimate that data centers could consume between 9.5% and 15.3% of total U.S. electricity by 2030, compared with 4.7% in 2024. The forecast represents a significant increase from earlier projections and highlights how accelerated AI adoption is reshaping the relationship between computing capacity and power availability. As hyperscale facilities, cloud platforms and AI workloads continue expanding, the challenge is no longer only building more generation but improving how existing energy systems respond. The next phase of data center growth may increasingly require grids to become more intelligent, flexible and efficient.

The debate around data center electricity demand is increasingly moving beyond simple consumption figures and toward the broader economics of grid management. Large computing facilities require consistent power availability, yet traditional electricity systems were designed around more predictable demand patterns from households and conventional industries. This creates pressure on utilities to expand infrastructure while ensuring that investment costs do not disproportionately affect existing customers. Recent analysis has shown that residential electricity prices remained broadly aligned with inflation between 2019 and 2024 before rising sharply last year. That shift has increased scrutiny over whether new industrial demand, including data centers, could accelerate cost pressures for ordinary electricity users. Therefore, the central question for policymakers is how to accommodate digital infrastructure growth without creating inefficient spending cycles.

Smarter Transmission Could Reduce the Cost of Data Center Expansion

One of the emerging solutions involves improving the performance of existing electricity networks instead of relying only on large-scale infrastructure expansion. Grid-enhancing technologies, including dynamic line ratings and advanced conductors, are gaining attention because they can increase transmission capacity without waiting years for entirely new systems. Researchers highlighted these technologies as commercially available tools that could provide near-term benefits by allowing utilities to operate existing networks more efficiently. According to an analysis cited by Columbia researchers from the Proceedings of the National Academy of Sciences, replacing existing transmission lines with advanced conductors could generate $180 billion in savings by 2050. The opportunity is significant because data center growth requires faster power delivery solutions while traditional transmission projects often face long development timelines. A smarter grid approach could allow electricity systems to absorb new demand while reducing unnecessary capital spending.

The financial challenge comes from the way utilities traditionally recover investment costs and earn returns. Under existing regulatory models, utilities generally receive returns on capital investments, which can encourage infrastructure expansion when demand increases. However, this model may not always prioritize the lowest-cost solutions for managing a changing electricity landscape. “This regulatory structure rewards capital deployment more than system optimization,” the researchers said. “As long as frameworks remain oriented toward proving ‘need’ for new capital investment rather than systematically evaluating lower-cost alternatives, the pace of deployment will lag.” This tension creates a strategic issue as data centers require more electricity but consumers remain concerned about rising bills. A future-focused grid model may need to reward operational efficiency alongside infrastructure development.

Demand Flexibility Could Become a New Tool for Managing AI Power Loads

Beyond transmission upgrades, demand response is emerging as another mechanism to manage the impact of large-scale computing demand. Data centers operate sophisticated workloads, some of which may offer flexibility under certain operating conditions. Instead of treating electricity demand as fixed, operators could shift certain workloads away from peak periods when grid pressure and prices are highest. The Columbia researchers emphasized that flexible data center operations could provide a valuable balancing mechanism during periods of intense electricity demand. “Large-load flexibility, particularly data center load shaping, or reducing usage during peak periods offers a complementary lever for managing peak demand without relying on high-cost, low-utilization generation,” the researchers said. This approach could allow AI infrastructure growth to support grid stability rather than becoming a source of additional strain.

However, implementing demand flexibility at scale requires changes in market structures, incentives and coordination between utilities and technology companies. Data center operators need clearer signals that encourage flexible operations, while utilities need frameworks that recognize efficiency as a valuable resource. In many regions, delays in connecting new generation and transmission infrastructure create gaps between rising demand and available supply. “Interconnection and new-build transmission timelines that take years sever the link between price signals and supply entry,” the researchers said. “Higher capacity prices mostly end up paying existing generators more rather than inducing new construction, and those extra costs are passed on to consumers through their electricity bills over time.” The issue reflects a broader challenge where electricity markets must adapt faster to the speed of digital infrastructure investment.

Utility Reform May Decide Whether AI Growth Raises or Lowers Costs

The long-term impact of AI data centers on electricity pricing will depend heavily on regulatory evolution. Performance-based regulation, which connects utility earnings to measurable outcomes, has been discussed as a possible alternative to traditional investment-driven models. Such approaches could encourage utilities to improve reliability, affordability and efficiency rather than simply expanding physical assets. However, researchers noted that evidence about the effectiveness of these models in reducing costs is still developing. The broader objective is creating a system where demand growth encourages better utilization of existing resources instead of triggering unnecessary infrastructure spending. As a result, the future of data center power economics could involve both additional capacity and improved coordination across the energy ecosystem.

Cost allocation will also remain a critical issue as utilities plan for rising electricity demand from digital infrastructure. Researchers pointed to the need for ensuring that data center-related expenses are distributed fairly while maintaining price signals that encourage additional supply development. If large electricity consumers receive benefits without carrying appropriate costs, existing customers could face greater financial pressure. Meanwhile, if pricing structures discourage investment, regions may struggle to attract new computing infrastructure. The balance requires policies that support economic growth while protecting affordability for households and businesses. Ultimately, the expansion of AI infrastructure will test whether electricity markets can become as adaptive as the technologies they power.

The Grid Opportunity Behind the AI Infrastructure Boom

The growing demand from AI-powered computing does not necessarily have to create a conflict between technology expansion and consumer affordability. A more advanced electricity network could transform data centers from passive energy consumers into active participants in grid management. Smarter transmission systems, flexible demand strategies and improved regulatory models could help unlock additional capacity without depending entirely on expensive new infrastructure. The researchers also highlighted the importance of addressing wider risks, including storms, wildfires and cybersecurity threats, which increasingly influence grid reliability. These challenges show that future electricity planning must consider resilience alongside capacity growth. The objective is building an energy system capable of supporting digital expansion while maintaining economic stability.

“Delivering affordable electricity depends on aligning supply, infrastructure, and cost allocation so that rising demand can drive higher utilization and lower costs — taking advantage of load growth as an opportunity, rather than a liability,” the researchers said. The statement captures the central challenge facing the power and technology sectors as AI adoption accelerates. Data centers will continue to require more electricity, but the answer may not be measured only in additional generation capacity. Instead, the next competitive advantage could come from creating smarter, more responsive grids that maximize every unit of power already available. As artificial intelligence reshapes digital infrastructure, electricity networks may become the next major platform requiring innovation.