A Data Center Isn’t Just a Plug-and-Play Solution

Welcome to Dispatch Energy! For many of us, electricity feels like a subscription service. We move in, contact the utility, establish an account, and expect the electrons to arrive on cue. The power system stands as one of the great coordination achievements of modern life precisely because, for most people, it fades into the background. When it functions smoothly, we barely notice it. When it falters, we notice nothing else.

That mental model does not scale. A data center seeking hundreds of megawatts is not simply another customer signing up for service. At that degree of scale, electricity provisioning becomes an infrastructure negotiation—and the challenge is not solely how much power data centers consume. The issue is how a large, fast-moving, uncertain new customer becomes a stable enough presence around which to structure power systems.

That distinction grows more crucial as artificial intelligence fuels a rapid expansion of data-center development and an expectation of further demand growth. Public discourse has centered on quantities: how much electricity data centers will use, how much new generation we will need, how much bills might rise. While those questions matter, their focus on electricity infrastructure shortages misses a deeper problem: a shortage of institutional infrastructure.

Large new electricity demand does not simply appear on the grid. It must be forecasted, studied, planned, financed, physically connected, and then delivered reliably, hour after hour. These steps require far more than engineering; they require credible commitments. 

A data center developer may seek the utility to reserve hundreds of megawatts of future capability, while the utility may need to construct substations, transmission upgrades, and transformers before the data center is fully operational. Regulators must then decide whether those investments are prudent, who should pay for them, and what occurs if the project is delayed, uses less electricity than forecast, or never materializes. The grid-owning utility must make lasting investments under uncertainty (and so does the hyperscaler). Better forecasting can reduce that uncertainty but cannot eliminate it. The harder institutional question is how to craft rules and contracts that enable investment without requiring ordinary customers to become involuntary insurers of speculative demand.

Data centers are not the first large customers utilities have ever served. The electricity system has a long history with energy-intensive industrial loads—steel mills, aluminum smelters, chemical plants, auto factories—that shaped the 20^th-century economy and the power grid. These clients received special contracts, industrial rates, and sometimes interruptible service in exchange for paying lower prices. Yet many historical large loads were tied to physical production processes, local resource bases, and long-lived industrial assets, arriving on timelines that aligned reasonably well with utility construction cycles.

Data centers disrupt that model by combining features that collectively strain the inherited utility planning process. They arrive in large, concentrated bursts, with a single campus demanding hundreds of megawatts and a few projects altering a utility’s load forecast. They are fast, with developers prioritizing “speed to power” while transmission upgrades, substations, and generation procurement proceed on slower clocks. They are mobile, shopping across utility territories and grid regions, with each site inquiry appearing as a potential future load before a developer commits to building anywhere. They are uncertain—some backed by major companies with real demand, others contingent on financing, permits, chips, or favorable tax treatment. And they may be flexible, though not automatically so, with some computing workloads more amenable to shifting across time than others.

A new paper on large-load tariffs and load forecasting, by Angela Navarro and Molly Knoll for the National Association of Regulatory Utility Commissioners, captures the key shift: Historical load forecast errors were driven by population growth, marginal changes in end-use consumption, and macroeconomic factors that were diffuse and incremental. Today’s large, discrete loads are localized, infrastructure-intensive, and often seeking service faster than utility planning cycles can accommodate. The old forecasting world concerned gradual demand trends; the new forecasting world centers on project-level shocks. Poor forecast accuracy becomes a central problem in the main character era.

For much of the 20^th century, load forecasting was necessary but relatively quiet, noticeable mainly when something went wrong because demand growth tended to be steady. The data-center boom changes that. In regions with substantial data-center interest, the load forecast becomes contested terrain.

The fundamental difficulty is the phantom-load problem. If developers request service in several locations before choosing a build site, utilities and regulators may see more proposed demand than will ever materialize. PJM, the grid operator for the mid-Atlantic, has responded by requiring firmer commitments for near-term forecast years while treating longer-term projects as less certain. ERCOT in Texas adjusted its large-load forecasting assumptions after noting an average project delay of 180 days and data centers’ consumption of less than half the capacity originally requested for study.

A forecast is easy to inflate when someone else bears the cost of relying on it. Forecasting is both technical and strategic, with developers wanting to reserve capacity, utilities potentially expanding their rate base, economic development officials courting investment, and consumer advocates all with a stake in the outcome. In such an environment, a load forecast becomes a claim on future infrastructure, and when forecasts prove wrong, the consequences are costly. Transformers, substations, and transmission upgrades cannot be returned with a simple receipt and a polite apology. Bad forecasts become bills.

This observation brings us to the central question of institutional economics: Who should bear the risk created by uncertain large-load forecasts?

The easy answer says that big technology firms should pay more because they are large and profitable. While politically persuasive, that framing is not precisely the right analytic lens. A better question is what commitments a large customer ought to make before others are asked to commit resources based on the forecast.

When a utility builds infrastructure to serve a large new customer, it makes durable, capital-intensive investments that are hard to repurpose. The relevant idea from transaction-cost economics is asset specificity: A transformer, substation, or transmission upgrade built to serve a particular large load in a specific location may hold little value if that load never materializes. This quality turns a forecasting problem into a contracting problem.

The framework is a classic setting for what are known as contractual hazards. One party must invest in durable assets before the future is fully known; the other party has better insight into its own plans and prospects, but it too faces durable investments. Once the investment is sunk, bargaining positions shift. MIT economist Paul Joskow’s research on coal contracts showed that contract duration tends to grow when relationship-specific investments are most consequential; when a party must invest in durable assets to serve another, longer contractual commitments lessen the risk of being left exposed. Nobel laureate Oliver Williamson described the need for “hostages” to ensure credible commitment: a deposit, collateral requirement, minimum bill, or termination fee gives the customer something at stake, making its forecast more than merely talk.

The practical question is: what kind of contingent contract can enable investment when the future remains uncertain? When parties cannot know the future with certainty, they can still agree in advance on how things will unfold under different future conditions. If the data center arrives on schedule, here is how costs will be recovered. If it progresses more slowly, here are the minimum charges. If it departs early, here is the exit fee. If it offers grid-stress flexibility, here is the compensation.

That is not punishment. It is governance.

Used properly, a large-load tariff becomes a rulebook for connecting to the grid when a customer is large enough to influence system planning, infrastructure needs, and how costs are allocated. These tariffs may require long-term contracts, minimum billing obligations, collateral, exit fees, phased ramp schedules, and notice requirements. The fundamental function is consistent: large-load tariffs transform forecasts into firm commitments.

The latest National Association of Regulatory Utility Commissioners paper adds a key insight. Tariffs are more than consumer protections—they can also improve forecasting. Minimum-demand obligations, collateral requirements, phased load ramps, eligibility thresholds, contract lengths, and exit fees all carry information about the timing, size, persistence, and credibility of large-load requests. Navarro and Knoll propose that regulatory commissions use tariff terms directly in load forecasting, applying them to baseline planning assumptions and sensitivity analyses, and updating forecasting assumptions as actual queue and realization data accumulate. This view shows that a large-load tariff is more than a price; it is a forecasting institution.

The elements of a well-designed tariff are not random regulatory add-ons. Long-term contracts address duration. Minimum bills handle downside demand risk. Deposits and collateral give the customer something at stake. Exit fees guard against costly early departure. Ramp and build schedules make timing uncertainty contractible. Curtailment provisions that define when the large-load customer will reduce demand specify what flexibility means in operational terms. The aim is not a flawless contract covering every possible future state; no tariff can foresee AI demand trajectories, chip constraints, or capital market movements perfectly. The more practical goal is contingent contracting: outlining what happens under the most important foreseeable deviations from the forecast.

The experience of AEP Ohio illustrates this well. After regulators approved and the utility adopted revised data-center tariff terms in July 2025—including 25-megawatt eligibility requirements, an 85 percent minimum billing obligation, 12-year contract terms, and termination fees equal to minimum charges for 36 months—the interconnection queue fell from 30 gigawatts to 13 gigawatts. The missing 17 gigawatts was not necessarily “phantom” in any simple sense. But once capacity requests became costly to maintain, the queue began to contain better information, filtering speculative entries before utilities built around them. Improved load forecasting can indicate which futures are more likely. Large-load tariffs help determine responsibility when the likely future proves wrong.

Requiring data centers to pay for the infrastructure they need is the necessary first principle. Existing customers should not subsidize speculative load growth. Yet stopping there misses the question of whether data centers could reduce the infrastructure they require by becoming more responsive to grid conditions.

Not all megawatts are created equal. A customer requiring firm, around-the-clock service imposes different costs from one able to reduce demand during shortages, shift computing tasks across time, or coordinate operations with grid needs. Some workloads—AI training, batch processing, cooling systems—may offer flexibility that latency-sensitive inference cannot. This flexibility is not automatically valuable, however. It must be defined, contractualized, and verified. A data center claiming flexibility should specify operational meaning: how much load can be reduced, how quickly, for how long, and at what compensation. Flexibility only proves valuable to the grid if it is reliable enough to plan around, making this also a credible-commitment problem.

Large-load tariffs could become tools for institutional discovery, revealing which customers are firm, which are flexible, and which are speculative—making their characteristics visible and actionable rather than forcing public utility regulators to guess.

The data-center boom did not expose regulatory weaknesses in electricity; it removed the luxury of ignoring them.

Forecasts are no longer mere appendices to utility plans; they sit at the center of economic development, reliability, affordability, and decarbonization. When a forecast becomes the basis for billions of dollars of infrastructure on both sides of a deal, the assumptions behind it deserve scrutiny: which projects are included; how credible they are; what commitments customers have actually made; and how much projected demand is firm, flexible, speculative, or duplicative.

But forecasting is only part of the answer. Forecasts reduce but do not eliminate uncertainty. Once we better understand the likely timing, size, and persistence of large new loads, we still need rules assigning responsibility when expectations prove wrong. Data centers are neither villains nor ordinary customers. They are large, sophisticated, fast-moving actors whose demand can reshape the electricity system, and good institutions should be able to distinguish between those generating real value and those claiming capacity without credible commitment.

Better forecasting tells us which futures are more probable. Better contracting tells us what happens when the actual future diverges. The first reduces uncertainty; the second governs the remaining uncertainty.

You cannot simply plug in a data center. But with improved rules, sharper forecasts, and stronger commitments, perhaps we can learn how to connect one.