Utilities Lead in Data Center Expansion
Utilities are investing heavily in data centers—what does this mean for the future of energy and infrastructure?
The power grid has a new most important customer: data centers. Once a niche concern for utility planners, they are now reshaping how major utilities think about capacity, capital allocation, and clean energy commitments. The utilities that move fastest aren't just keeping the lights on; they're becoming critical infrastructure partners in the AI economy.
Duke Energy, Northern Indiana Public Service (NIPSCO), and Xcel Energy are among the utilities already fielding major data center projects in their service territories, according to testimony the Edison Electric Institute delivered to federal regulators. That's not a footnote; that's a signal about where load growth is concentrating and which players are positioned to capitalize on it.
Why Data Centers Are Consuming the Grid
The numbers help explain the urgency. A single hyperscale data center can draw 100 MW or more of continuous load — roughly equivalent to powering 80,000 homes. Now multiply that across dozens of campuses, and the grid math gets complicated fast. AI workloads are pushing that ceiling higher. Training large language models requires sustained, intensive compute that doesn't cycle down the way commercial or residential load does; it runs constantly.
This isn't a demand spike utilities can plan around with seasonal forecasting — it's a structural shift in baseline load.
Historically, utilities built capacity models around gradual, predictable growth: population changes, industrial cycles, and the electrification of heating. Data center demand breaks that model. A single signed lease agreement with a hyperscaler can add more load to a substation than a decade of organic residential growth. Utilities and their regulators are scrambling to adapt procurement timelines, interconnection queues, and rate structures to match the pace.
The geographic concentration matters too. Data centers cluster around specific corridors — Northern Virginia, Phoenix, Dallas, Central Ohio — drawn by land costs, fiber density, water availability, and favorable regulatory environments. When those clusters sit inside a utility's service territory, the implications ripple through every planning document that utility produces.
Who's Actually Doing the Work
Duke Energy's footprint across the Carolinas and the Midwest has made it a natural landing zone for data center development. The company has been vocal about its expectations for significant load growth driven by technology customers, and it's been adjusting its integrated resource planning accordingly — meaning more generation capacity, more transmission investment, and faster interconnection timelines.
NIPSCO's position is arguably more interesting. Northern Indiana isn't traditionally a top-tier data center market, but NIPSCO has been navigating a high-profile coal retirement and clean energy transition while simultaneously managing new large-load inquiries. Attracting data center load in that context is a double-edged proposition: it provides revenue stability that makes the transition financially viable, but it also requires ensuring the new load can be served cleanly.
Xcel Energy, serving Colorado and Minnesota among other states, brings its own complexity. The company has long-standing renewable energy commitments and operates in states with aggressive clean energy standards. Data center utility projects in Xcel's territory don't just have to work economically — they have to work within a framework where carbon matters to regulators and customers.
The Edison Electric Institute flagging these utilities specifically in federal testimony isn't a coincidence. EEI advocates for the industry before FERC and other federal bodies, and presenting concrete examples of utility-scale data center engagement signals that the sector wants regulatory frameworks that accommodate — rather than bottleneck — this load growth.
The Financial Stakes
Here's where stakeholders need to pay attention. Data centers are, from a utility perspective, nearly ideal customers. They're creditworthy, consume enormous quantities of power at predictable rates, and often sign long-term agreements that provide the kind of revenue visibility utilities use to justify capital investment. A utility that locks in a 15-year power supply agreement with a major hyperscaler has, in effect, de-risked a significant chunk of its generation portfolio.
That de-risking has real consequences for infrastructure investment. When utilities can demonstrate contracted load, they can make the case to regulators for rate-based capital expenditures — new transmission lines, upgraded substations, expanded generation. Ratepayers ultimately bear that cost, which is where the politics get complicated, but the investment itself creates tangible grid improvements that benefit the broader service territory.
The risk isn't that utilities invest too much in data center infrastructure — it's that the rapid pace of data center development outstrips the utility's ability to interconnect and serve that load reliably.
Interconnection queues in many regions are measured in years, not months. A data center developer who signs a lease expecting power delivery in 18 months may find themselves waiting three years for a grid connection. Utilities that can compress that timeline — through pre-permitted substations, pre-built transmission capacity, or streamlined large-load processes — have a genuine competitive advantage in attracting data center utility projects to their territories.
The Clean Energy Complication
Data centers and clean energy exist in a genuinely complicated relationship. On one hand, major technology companies — Google, Microsoft, Amazon, Meta — have made highly public commitments to 100% renewable energy and net-zero carbon operations. That creates demand pressure on utilities to deliver clean power, not just any power. On the other hand, the sheer volume of load that AI-era data centers represent is straining grid systems that are simultaneously retiring fossil generation and waiting on new renewable capacity.
The gap between aspiration and reality can be significant. A hyperscaler might sign a virtual power purchase agreement for wind energy in one state while taking power from a grid that's still 40% coal in another. The accounting works; the electrons don't always match the story.
Utilities like Xcel, with existing renewable portfolios and ambitious clean energy timelines, are better positioned to offer data center customers genuine carbon-free power. But even Xcel faces constraints. Renewable intermittency means data centers — which need 24/7 reliable power — require backup capacity. Battery storage is increasingly filling that role, but the economics and availability of grid-scale storage remain a constraint, particularly for the largest loads.
What this creates, practically, is a tiered market. Data centers with genuine clean energy requirements will gravitate toward utilities with credible renewable portfolios, creating a sustainability premium for grid operators who have done the hard work of decarbonizing.
NIPSCO's ongoing transition from coal to renewables is, from this lens, not just an environmental story; it's a customer acquisition strategy for the next decade of data center growth.
What Comes Next
The regulatory environment around data center load growth is still forming. FERC has been examining interconnection reform, and state public utility commissions are increasingly being asked to weigh in on how large-load customers should be treated in rate design. Should data centers pay higher rates to reflect the infrastructure cost they impose? Should they be required to source a minimum percentage of renewable energy? Should utilities be permitted to accelerate capital deployment outside of normal rate case cycles when large-load customers are involved?
These aren't hypothetical questions. They're active proceedings in multiple states, and the answers will shape the economics of data center utility projects for years.
For developers and investors, the practical takeaway is straightforward: service territory matters more than it used to. A data center site in a jurisdiction with a utility that has pre-built interconnection capacity, a credible clean energy portfolio, and a constructive regulatory relationship is worth more than a cheaper piece of land served by a utility still working through coal retirements with an overloaded interconnection queue.
For the utilities themselves, the strategic imperative is to stop treating data center load as an opportunity and start treating it as a core planning assumption. The utilities that will win the next decade of infrastructure investment — Duke, NIPSCO, Xcel, and others positioned like them — are the ones building the capacity, the processes, and the regulatory relationships now, before the demand fully arrives.
The grid is being rewired around a new anchor tenant. The utilities that understand that earliest will be the ones writing the terms.
Call to Action: Explore how InfraSale can help you navigate the evolving landscape of data center utility projects. Visit InfraSale Marketplace for more insights.
[INTERNAL LINK: data center growth]
[INTERNAL LINK: clean energy commitments]
[INTERNAL LINK: utility infrastructure investment]