Edward Burgess is a partner at Current Energy Group.
The debate over the impacts of new data centers on customers' electricity bills is only intensifying.
On the one hand, many people blame data center companies for triggering new grid costs that others will pay. On the other, data centers hold great promise to lower customer bills.
For the latter, the theory is that, as high-utilization infrastructure users, data centers can absorb a large share of fixed costs that would otherwise be charged to existing customers.
To their credit, nearly 200 data center companies have signed the Ratepayer Protection Pledge, committing to pay their fair share of costs. However, achieving this commitment depends on cooperative action by local utilities to update their rates accordingly.
This raises an important question: Are investor-owned utilities willing to share the benefits of large-load growth with existing customers? Recent evidence suggests otherwise in some cases.
To fully understand this problem, let’s revisit the fundamentals of electricity ratemaking.
New rates are typically proposed by utilities but ultimately approved by state utility regulators. When utilities propose new rates, the core calculation is based on two key inputs: 1) utility costs in the numerator, and 2) utility sales in the denominator.
If the denominator increases more than the numerator (i.e., sales increase more than costs), then customer rates should decline on average.
But in the era of hyperscale load growth, the specifics of these inputs — how they are defined and when relative changes occur — matter a great deal. In several recent cases, utilities have simply neglected to increase kWh sales from new data center load in their rate proposals, while still including related cost increases.
For example, in North Carolina, Duke Energy recently proposed increasing customer rates by over 15% over the next two years.
However, this increase was based on a rate calculation that kept the denominator essentially frozen at 2024 levels — even though Duke had already contracted to add hundreds of MWs of new data centers in the coming years, and included costs triggered by this in its proposal.
Current Energy Group analyzed Duke’s proposal and presented expert testimony to the North Carolina Utilities Commission on behalf of the state’s Attorney General’s office. We found that 54% of the requested rate increase — approximately $1 billion annually — could be erased if kWh sales in the rate formula were simply corrected to reflect Duke’s own planning assumptions.
In fact, correcting just the kWh sales figure and one other key variable — return on equity — was enough to turn the proposed 15% rate increase into a 4% decrease, even with significant new grid investments.
A similar pattern emerged in Missouri, where Evergy Missouri Metro included costs for new gas generation to serve new data centers but did not account for the increased sales to those data centers in its rate calculation. As data center sales ramp up, the utility would see a windfall over the four years the new rates are in effect.
In Louisiana, increased sales from a new Meta data center could trigger an immediate customer refund over the next three years. Instead, Entergy proposed accounting methods to defer that benefit over a 20-year period.
This “one simple trick” of immediately updating kWh sales in the rate formula could save customers billions in the coming years. So why do utility rate proposals keep failing to account for data center load growth?
Some utility analysts might point to arcane explanations, such as the traditional practice of using a “historic test year.” There may be a grain of truth to that, but it is not a convincing argument given widespread concerns about affordability.
Duke, for instance, had the legal ability to propose a forward-looking Multi-Year Rate Plan but chose not to include any load growth in its rate proposal.
A more cynical view is that utilities are seeking to “lock in” higher rates before new data center load fully materializes and the denominator rises. This creates a major windfall for utilities, at least temporarily, as data center load ramps up before rates can catch up.
Traditionally, utilities have complained about “regulatory lag,” which arises when rates are not increased quickly enough to capture rising costs. Now, in the era of hyperscale load growth, we have the opposite problem — a “reverse regulatory lag.” Instead, rates are not being reduced quickly enough to capture rising revenue from large loads.
This problem is exacerbated by another challenge: As utilities slow-walk updates to the denominator (kWh sales), they speed-run updates to the numerator (costs recovered).
Many utilities, including Duke, use accelerated cost recovery mechanisms, such as Construction Work in Progress, to fund new infrastructure for load growth.
This further strains affordability. Infrastructure costs are front-loaded through near-term rate increases, while any rate relief from load growth won’t materialize for years.
Under Duke’s original proposal, customers would pay for the construction of new gas-fired power plants four years before they start operating, and six years before data center load appears in the denominator.
As affordability concerns persist, what solutions should regulators consider?
- Regulators should insist on up-to-date sales forecasts for any rate proposal and apply them immediately. If utilities rely on forecasts to plan and build generation, they should rely on those same forecasts to reduce rates by updating the billing determinants.
- When revenue grows rapidly, regulators should ensure that utilities absorb more of the initial cost and risk involved rather than passing this on to consumers. For example, regulators should consider halting accelerated cost-recovery mechanisms. The regulatory lag problems those mechanisms sought to address are much less salient now.
- For certain states, full revenue decoupling mechanisms can offer a solution. However, this might not be available in all cases or benefit all customer segments.
Whether hyperscale load growth raises or lowers bills depends on how and when the numerator and denominator are updated. Understanding these mechanics and the options for mitigating them will shape how regulators and stakeholders weigh the bill impacts of large load growth.
