Is Growing Power Demand Reshaping Utility Planning?

The numbers are hard to ignore. U.S. data center power consumption is projected to reach 35 gigawatts by 2030—roughly equivalent to the entire electricity consumption of California today. Utilities that were planning for 1–2% annual load growth are suddenly staring down requests for 500 MW interconnections from single customers. That's not an incremental shift; that's a structural break.

For decades, utility planning was a relatively predictable discipline. Demand grew slowly, forecasting models were well-calibrated, and the biggest surprises were weather-related. That era is over. The convergence of hyperscale data centers, EV adoption, onshoring of manufacturing, and cryptocurrency mining has created a demand environment that legacy planning frameworks simply weren't built to handle.

The Demand Surge Is Real—and It's Concentrated

Data centers don't spread load evenly across the grid the way residential customers do; they cluster. Northern Virginia—already home to the world's largest concentration of data centers—added more than 2,700 MW of new capacity in 2023 alone. Similar buildouts are accelerating in Texas, Georgia, Arizona, and the Pacific Northwest, creating intense pressure on specific transmission corridors and substations.

The challenge isn't just total load growth—it's where that load lands and how fast it ramps.

A hyperscale facility can go from groundbreaking to energization in 18–24 months. Traditional utility infrastructure projects—new transmission lines, substation upgrades, generation capacity additions—routinely take 5–10 years to complete. That gap between customer timelines and utility delivery timelines is where the real crisis lives.

Energy-intensive industries beyond data centers are compounding the problem. Semiconductor fabs like TSMC's Arizona facility require hundreds of megawatts of highly reliable power. Green hydrogen electrolysis projects need gigawatts. EV charging infrastructure, while more distributed, still represents significant incremental load on distribution systems that were designed for a different era.

How Utilities Are Adapting—Some More Successfully Than Others

The utilities that are getting ahead of this aren't doing it through incremental improvements to their old planning processes. They're rebuilding the process from scratch in a few key ways.

Probabilistic load forecasting is replacing deterministic forecasting at forward-thinking utilities. Instead of predicting a single demand curve, planners now model multiple scenarios weighted by probability—accounting for the chance that three proposed data campuses in their territory actually get built versus one or two. Dominion Energy, which serves the data center-dense Northern Virginia market, has essentially created a dedicated large-load interconnection team to manage this pipeline.

Proactive infrastructure investment is the other major shift. Historically, utilities built infrastructure in response to confirmed load—a customer signs an agreement, then construction begins. The new model at utilities like Georgia Power involves investing in "ready sites"—substations and transmission capacity that can support large loads before a specific customer commits. It's a speculative investment, but the alternative is telling a Fortune 500 company they need to wait seven years for power.

Demand response and flexible load agreements are also gaining traction. Some data center operators, particularly those running less latency-sensitive workloads like AI training, are willing to curtail consumption during grid stress events in exchange for rate benefits. These agreements give utilities a real-time buffer that buys time for longer-term infrastructure solutions.

The utilities that are struggling tend to share a common trait: they underestimated the pace of the buildout and are now managing a queue of interconnection requests without the staffing, capital, or regulatory authority to process them efficiently.

The Financial Stakes for Utilities That Don't Adapt

Utilities that fail to keep pace with large-load demand aren't just leaving growth on the table—they're creating financial and regulatory exposure.

When power delivery fails to materialize on promised timelines, the consequences cascade. Data center developers who can't get power commit their capital elsewhere. Economic development projects relocate to other states. Regulators start asking uncomfortable questions about planning competence. And in deregulated markets, where competitive dynamics are sharper, utilities can lose anchor customers to regions with faster interconnection queues.

The inverse is also true: utilities that position themselves as reliable partners for energy-intensive industries can command significant long-term load commitments. A single hyperscale data center represents decades of predictable, high-density load—exactly what a utility needs to justify infrastructure investment to regulators and bond markets.

Energy prices are another pressure point. Rapid, unplanned load growth strains grid resources in ways that drive up costs for all ratepayers. If a utility adds 2,000 MW of new data center load without corresponding generation and transmission additions, the resulting congestion and reliability costs get socialized across the rate base. That creates political and regulatory friction—the kind that can delay future rate cases and infrastructure approvals.

Long-Term Strategies and the Technologies Reshaping the Equation

Several technologies are changing what's possible for utility planners operating in this environment.

Grid-scale battery storage is enabling utilities to defer traditional infrastructure investments by providing capacity at critical chokepoints. A 200 MW / 800 MWh battery installation at a constrained substation can bridge the gap while transmission upgrades work through permitting. The economics have improved dramatically—battery storage costs have dropped roughly 90% over the past decade—making this a legitimate planning tool rather than a demonstration project.

Advanced conductor technologies, particularly high-temperature low-sag (HTLS) conductors, allow utilities to increase the capacity of existing transmission lines without building new structures. In corridors where new right-of-way is unavailable, this can be a critical unlocking technology.

Behind-the-meter generation is shifting the dynamic for some large energy users. Data center developers are increasingly pursuing co-located solar, fuel cells, and even small modular nuclear reactors to reduce their dependence on grid interconnection. Microsoft's agreement to restart a Three Mile Island unit specifically to power its data centers signals that the largest operators are willing to go further than most expected to secure reliable power.

For utilities, this creates both opportunity and risk. On-site generation reduces interconnection demand, which eases grid planning. But it also reduces the revenue base that utilities depend on to fund grid infrastructure— infrastructure that still needs to be maintained for resilience even when customers generate their own power.

What Industry Professionals Should Be Doing Now

If you're on the development side—whether you're building data centers, industrial facilities, or energy storage projects—the single most important thing you can do is start utility conversations earlier than feels necessary. If you think you need to talk to the utility when you have a signed lease, you're already behind. The developers getting power on schedule are the ones who engaged utilities at the site selection stage and understood the interconnection queue dynamics before they committed capital.

Site selection itself is increasingly a utility planning exercise. A parcel 10 miles from a constrained substation might be cheaper to acquire, but if it adds 18 months and $40 million in upgrade costs to your interconnection timeline, the math inverts quickly. Proximity to existing high-voltage infrastructure, substation capacity headroom, and local utility responsiveness are first-order site selection criteria—not afterthoughts.

For utilities and their planning teams, the lesson from utilities that have navigated this well is straightforward: treat large-load pipeline development as a partnership, not a transaction. The data center developers who are choosing locations are choosing partners as much as they're choosing geography. Utilities that invest in dedicated large-customer teams, transparent interconnection processes, and proactive infrastructure development will attract a disproportionate share of this buildout.

The load growth happening right now isn't a temporary spike. AI infrastructure, domestic manufacturing reinvestment, and electrification are structural trends with decade-long runways. Utility planning that's calibrated to the demand environment of 2010 isn't just outdated—it's a liability. The utilities rewriting their planning frameworks today are building the competitive position that will define their next 20 years.

[INTERNAL LINK: utility planning strategies]

[INTERNAL LINK: energy-intensive industries]

[INTERNAL LINK: data center demand trends]

EDITOR NOTES

• Consider cutting the paragraph discussing the common traits of struggling utilities, as it may feel repetitive.
• Ensure that the internal links are relevant and lead to appropriate content on the blog.