Episode 1 – A Whole Lotta Load: The Next Wave of Utility Demand

Marc (00:42)
Today’s episode is titled A Whole Lot of Load: The Next Wave of Utility Demand. For years, utilities were planning around relatively stable demand growth, but suddenly that’s changing. Data centers, AI infrastructure, advanced manufacturing, and industrial reshoring are driving a new wave of load growth that is larger, faster, and more concentrated than what most utilities have dealt with in decades.

And while these projects create major economic development opportunities, they also introduce significant challenges around planning, infrastructure, risk, and execution. So the question becomes how do utilities prepare for this next era of demand growth? To explore that, I’m joined today by Chris Sturgill, who works closely with utilities on large load strategy planning and customer engagement. Chris, we like to start with.

Big picture. Utilities have dealt with low growth before, whether it was air conditioning, industrial expansion, or broader economic growth. What makes this wave fundamentally different?

Chris (01:54)
Hey, thanks, Marc, and happy to join you today. I think what makes this wave different is the combination of scale, speed, concentration, and uncertainty. So, as you mentioned, historically, load growth was gradual and broadly distributed across the residential, commercial, and industrial customers. And thinking about that shift to air conditioning, that load was diffuse across an existing customer base, and it took decades to become fully saturated. Today we’re seeing individual projects

that request hundreds of megawatts or sometimes even approaching a gigawatt of power from a single site. At the same time, AI hyper AI and hyperscale data center developers are looking for power in one to three years, while the infrastructure required to serve that tremendous load often takes much longer to build. And so utilities are being asked to make decisions on unprecedented levels of load with much less certainty than they’re accustomed to.

Marc (02:50)
Is the biggest challenge the amount of load itself or the pace and uncertainty around that load? Because it feels like utilities are suddenly being asked to make very large infrastructure decisions on much shorter timelines.

Chris (03:07)
Yeah, I’d I’d say it’s the pace and uncertainty coupled with the scale is the biggest challenge. Utilities know how to build infrastructure when that demand is certain. The problem is that many projects are still evolving, and utilities have to make these billion-dollar decisions years before the load actually materializes. And managing that uncertainty while moving fast enough to remain competitive is really the balancing act at hand.

Marc (03:30)
And that creates a very different economic development conversation. Utilities and states are competing very aggressively for these projects, and in what we’ve seen, but what makes a utility service territory attractive to large load customers like data centers?

Chris (03:49)
Yeah, the first thing customers look for is available power and a credible path to obtaining it quickly. Beyond that, they look for things like transmission access, reliability, resiliency and natural disaster considerations, electricity costs, fiber connectivity, land availability, permitting timelines and local policy support. So it’s a whole gamut of considerations. Some of those factors are within the utility’s control, while others are not.

In the land rush that we’ve seen for selecting locations, a lot of the low-hanging fruit or areas with existing capacity that check all those boxes are gone, which means putting together compelling offerings where the criteria is met, and where there are gaps, plans to close those quickly.

Marc (04:35)
Are we starting to see utilities think about economic development differently because of these opportunities?

Chris (04:42)
Absolutely. I think historically economic development was viewed as a supplemental activity. And today for many utilities, large load attraction has become a strategic priority. Utilities are proactively planning infrastructure, developing specialized tariffs, and coordinating with state and local economic development agencies and building dedicated teams all focused on this question of serving data centers and large load customers.

Marc (05:09)
Utilities have traditionally viewed all load growth as purely positive. Are utilities becoming more selective about the types of large load customers they want to attract, or is all load growth still considered good growth?

Chris (05:25)
Yeah, that’s right. Historically, everyone did view any load growth as positive because growth was relatively predictable and the infrastructure needed to serve it was incremental. Today, a single customer may require hundreds of megawatts of new capacity and billions of dollars of supporting infrastructure, which changes the economics and the risk profile significantly. So utilities are increasingly evaluating not just the size of the load, but also the quality of it. They’re asking questions like how certain is the project?

What’s the customer’s financial strength? How quickly will the load ramp? And what’s the expected load factor? Will the customer provide any load flexibility? And how much inf infrastructure do we need to invest in relative to the long-term value that’s created? In many ways, the conversation has shifted from how do we attract as much load as possible to how do we attract the right load? Utilities still want growth, but they’re looking for projects that align with their planning horizon.

And create durable economic value and can be served without creating undue risk for existing customers.

Marc (06:27)
But attracting that load is only part of the challenge; then you have to actually build the infrastructure to serve it. Given that serving these customers can require enormous investments in generation, transmission, and distribution infrastructure, how are utilities thinking about financing that build-out?

Chris (06:48)
The scale of investment is enormous. Utilities may need to build their generation, transmission, substation, and distribution facilities years before they ever see revenue from these customers. So they’re looking at options to finance this transformative build-out. For instance, Duke Energy and Excel Energy are both looking at using private credit to support their CapEx plans. And what private credit offers is more flexibility in terms of the duration and collateral that they need for the

For the financing terms, and can let utilities ring fence the investments for, say, a large data center capacity expansion.

From the rest of their ratepayers.

Marc (07:28)
One of the major concerns utilities have is stranded asset risk or counterparty risk, because they’re building this large infrastructure to meet significant new demand. How are utilities protecting themselves against that risk that projected load doesn’t fully materialize?

Chris (07:50)
Yeah, we’re seeing a number of ways that utilities are looking at risk management. They’re looking at strengthening commitment mechanisms through deposits, collateral requirements, minimal demand charges, long-term contracts, exit fees, and stranded cost recovery provisions. They’re also looking at phasing infrastructure investments so that additional capital is deployed only when customers meet specific development milestones. The idea here is to align investment timing with the customer commitment levels.

Marc (08:21)
Are regulators generally supportive of these investments or are they becoming more cautious given the scale and uncertainty? Because in some cases utilities are being asked to spend billions before the load is fully online.

Chris (08:38)
Yeah, I think regulators are recognizing the economic development opportunities and the ability to spread costs over a greater number of megawatt hours, but they’re becoming increasingly protective on looking after their existing customers. The key question is whether investments are being made prudently and whether the costs are being allocated appropriately. Regulators are asking utilities to demonstrate that the current set of ratepayers won’t be left holding the bill.

If these projects don’t proceed as expected.

Marc (09:10)
Even if the economics work, utilities still have to solve the planning problem created by that uncertainty. How are these large loads challenging traditional utility planning and forecasting processes?

Chris (09:25)
Traditional forecasting models have relied heavily on historical trends. Data centers don’t fit that model. When a single project can be larger than years of normal load growth, utilities must increasingly supplement that traditional forecasting with discrete project forecast and scenario planning and probabilistic approaches that account for different levels of customer certainty.

Marc (09:47)
One thing we hear constantly is the mismatch between data center development timelines and utility infrastructure timelines. How significant is that challenge? Because the customer may want power in two or three years, but major infrastructure can take much longer.

Chris (10:08)
It’s one of the defining challenges in the industry right now.

As you noted, the new infrastructure for transmission or generation projects can take a decade or more. That mismatch puts tremendous pressure on utilities to find creative solutions while maintaining reliability. A good example of what’s happening in Northern Virginia is where Dominion is managing some of the largest data center growth in the world.

One of the approaches they’re using is to deliver power in increments, as most data centers don’t need all of their requested power right away. It started that practice in Loudoun County, which is known as Data Center Alley, in 2022, but it’s expanded that strategy to other areas. And the approach is enabled by a rate structure that phases in costs and demand floors to match the pace of the data center development and the related infrastructure, while protecting against cross-subsidization for other customers.

Marc (11:05)
And how are utilities adapting their planning processes to manage that mismatch that you talked about?

Chris (11:13)
It starts at the outset by formalizing intake and interconnection processes, and it creates stage gates that are tied to the customer commitments. From there, it improves coordination across planning, engineering, regulatory, and customer teams. Utilities are also evaluating phased energization approaches, like we just discussed in the Dominion example, scenario-based planning, and alternative supply arrangements to better manage uncertainty.

Marc (11:39)
Yes, I’ve definitely seen that in our work in the industry. Of course, these projects also create broader stakeholder and community conversations because when projects move this quickly, stakeholder engagement becomes critically important. What are the biggest stakeholder concerns utilities are navigating when these projects come into a service territory?

Chris (12:03)
The most prominent concern right now is affordability. Stakeholders want to know whether existing customers will pay for infrastructure built to serve data centers. Beyond that, there’s concerns about reliability, water consumption, land noise, construction impacts, and whether communities are receiving sufficient benefits for the investment.

Marc (12:23)
A lot of these conversations ultimately show up in tariffs and rate design. How are utilities approaching cost allocation and risk sharing for these large load customers like data centers?

Chris (12:37)
The principle underpinning this is cost causer pays, and that philosophy is pretty straightforward. Customers should pay for the costs they create. When those costs are fully incremental, that math is straightforward. But there’s a key conversation about how the existing infrastructure that data centers are using should be valued and how those costs should be recovered. One of the biggest tensions today is determining where that line sits between infrastructure built specifically for a data center.

And infrastructure that creates a broader system value. Data center developers, understandably, don’t want to pay for assets that will eventually benefit future customers. On the other hand, regulators and consumer advocates want assurance that the existing customers aren’t subsidizing speculative growth. Another area of contention is risk allocation. Utilities are being asked to make large investments before the load’s fully online, while developers want certainty around pricing and timelines.

Regulators are increasingly scrutinizing whether the utility has adequate protections in place if a project’s delayed, downsized, or even canceled. AEP Ohio’s data center tariff, for instance, includes long-term minimum payment obligations that require the customers to pay for most of the capacity they reserve, even if they don’t fully use it. Georgia Power has implemented special contract frameworks for very large loads, and we’re beginning to see utilities explore flexible service models.

Where customers can receive faster or lower-cost service in exchange for accepting some operational constraints. The common theme is that utilities are trying to better align their customer commitments, the infrastructure investment, and risk allocation.

Marc (14:15)
Even when utilities align on planning, tariffs, and economics, there’s still the challenge everyone talks about now: speed to power. That’s become one of the defining issues in this space. What does that phrase really mean in practice?

Chris (14:31)
At its core, speed to power is the ability to deliver reliable electricity on a timeline that matches customer expectations. For a data center developer, time is money. The AI race is on, and more quickly securing power and beginning operations can represent a competitive advantage. Utilities that can accelerate those timelines would differentiate themselves as an energy infrastructure partner.

Marc (14:55)
Where are the biggest bottlenecks when utilities try to accelerate those timelines?

Chris (15:02)
In short, everywhere. The bottlenecks span the entire value chain. Transmission development, generation construction, permitting, interconnection studies, transformer availability, switchgear procurement, and workforce constraints all can contribute. In many cases, even when the funding is available, supply chain limitations and permitting timelines become the critical path.

Marc (15:24)
What options are utilities exploring to shorten those timelines or create interim solutions to address the need?

Chris (15:34)
They’re looking at a couple of options. As we mentioned with the Dominion example, they’re looking at phased energization strategies, temporary generation, on-site generation or co-location agreements, flexible service agreements, and proactive investment in infrastructure in high growth areas. Again, anything that can bridge the timeline between what the customer is expecting and how long it takes to build these assets.

Marc (15:59)
Chris, one concept that comes up more and more in these conversations as utilities try to manage these challenges is load flexibility. What do we mean when we talk about load flexibility for data centers and other large load customers?

Chris (16:17)
I think it goes back to defining our terms. When people hear load flexibility, they often think about traditional demand response programs where customers reduce their usage during peak periods. But what’s emerging now is really broader than that. Load flexibility is the ability of a large customer to adjust how and when it consumes power in response to grid conditions. This could mean temporarily reducing load, shifting some comp computing workloads to another time or location.

Utilizing on-site generation or storage, or operating under service agreements that allow some level of curtailment during system stress.

What’s new is that AI data centers may actually be more flexible than people assume. Recent work from Duke University’s Nicholas Institute suggests that even modest levels of flexibility could unlock significant capacity on the existing grid. Their research introduced the concept of curtailment-enabled headroom, essentially asking how much additional load could be added if customers were willing to accept very limited interruptions or adjustments in operations. The study found that nearly 100 gigawatts of new load could be.

Potentially integrated across major US balancing authorities with relatively small amounts of annual curtailment. The interesting question isn’t whether these options are technically possible. The real question is whether data center operators have the appetite to accept those trade offs. Historically, the answer would have been largely no. Data centers have been built around extremely high reliability standards, and operators have been reluctant to accept anything that could jeopardize uptime. But AI may be changing that equation.

Research from Goldman Sachs highlights that many AI training workloads are fundamentally more flexible than traditional cloud workloads. Unlike applications that require real-time responsiveness, AI training jobs can often be paused, delayed, or shifted geographically with limited impact. Goldman offers that this flexibility could allow AI data centers to participate in curtailment programs, where they operate at full capacity most of the year, but temporarily reduce consumption during periods of grid stress.

That doesn’t mean that hyperscalers are suddenly eager to be interrupted. What it means is that they’re increasingly evaluating the trade-off. In some markets, the choice isn’t between firm service and flexible service. It’s between flexible service today against firm service several years from now. As utilities struggle to build generation and transmission quickly enough, that becomes a much more compelling business discussion. That speed-to-power trade-off and the value at stake in achieving AI milestones first.

Make it likely that someone will try to make this work. Whether or not that reaches mainstream adoption will depend on how far utilities and data centers are willing to push beyond traditional approaches to find common ground. I don’t think it’ll become a default option because firm service is so much easier to manage and plan for all parties, but I expect many will reach for it out of necessity in the next decade.

Marc (19:12)
Aren’t both utilities and customers still somewhat hesitant to fully embrace it?

Chris (19:18)
They are, and I think the biggest issue is operational risk. Data center operators are highly focused on uptime and reliability, and utilities need confidence that flexibility commitments will be performed when needed, especially when it’s hundreds of megawatts at hand. Developing commercial structures, operational protocols, and contractual agreements that satisfy both parties is complex and still evolving.

Marc (19:42)
Chris, when you look ahead, what do you think will differentiate utilities that successfully manage this next wave of demand growth from those that struggle?

Chris (19:53)
I think the winners will be utilities that can balance the growth, risk, and execution. Those would probably look like utilities with strong planning processes, clear customer engagement strategies, disciplined risk management, effective stakeholder communication, and the organizational ability to move quickly. Success won’t be defined simply by attracting the load. It will be defined by serving that load sustainably and responsibly.

Marc (20:19)
Thanks, Chris. That makes sense. If there’s one thing utility leaders ought to be thinking about right now when it comes to large low growth, what would it be?

Chris (20:30)
Utility leaders should focus on building a repeatable framework rather than reacting to individual projects. The utilities that succeed will establish clear policies around planning, customer commitments, cost allocation, risk management, flexibility, and speed to power. The volume of opportunities is only increasing, and having a scalable approach will become more valuable than solving each project one at a time.

Marc (20:56)
Thank you, Chris, and thank you for joining me today. I appreciate you sharing your perspective on what utilities are facing as this next wave of demand growth accelerates.