The Affordability Puzzle: Separating Utility Costs from Customer Perceptions

Michael Chua
PowerShift Contributor

First Published Q3 2026

On Aug. 3, Gov. Greg Abbott directed the Electric Reliability Council of Texas (ERCOT) and the Public Utility Commission (PUC) of Texas to audit all data center projects moving through the interconnection process. Verify the load. Confirm that the megawatts on paper correspond to something somebody actually intends to build. Approvals, in practice, stopped.

This is in Texas. This is the grid that prides itself on connecting things faster than anybody else, in a state whose economic development pitch, “Bigger. Better. Texas.”, fits on a bumper sticker. Two months prior, Abbott had directed the PUC to require data centers to carry the full cost of the transmission lines, substations, and upgrades built to serve them. But why the abrupt turn around?

Texas did not lack gas, land, capital, turbines, or ambition. It ran out of public consent, a challenge now emerging nationwide.

The Queue We Talk About, And The One We Don’t

Ask a planner what’s holding up the buildout, and you’ll get an answer about the queue. In its 2026 edition of “Queued Up,” Lawrence Berkeley National Laboratory counted 2,060 gigawatts of generation and storage actively seeking U.S. interconnection at the end of 2025, spread across roughly 8,200 projects. Another 549 GW already hold a draft or executed interconnection agreement and still aren’t running. Globally, the International Energy Agency estimates that more than 2,500 GW of renewables, storage, and large loads are stuck in connection queues and says annual grid investment must climb about 50% from today’s $400 billion to meet demand through 2030.

That is the engineering perspective, familiar to industry professionals. However, there’s another holdup that’s taking root and growing fast: the opposition of the public.

In the first quarter of 2026 alone, more than 75 data center projects worth roughly $130 billion were blocked or delayed by local opposition, according to Data Center Watch. That’s about what the whole of 2025 produced, compressed into three months. The number of active opposition groups more than doubled to 833, in 49 states. At least 69 local governments have passed outright bans (Canary Media: June 2026). Seattle, home to two of the largest hyperscalers on earth, adopted a one-year pause affecting five proposed projects (McNichols: June 2026). Denver’s city council voted unanimously for its own moratorium (Denverite: May 2026). And in New York, the legislature passed a statewide one-year moratorium on new permits for facilities above 20 megawatts (DLA Piper: February 2026). The list goes on, with the likes of Oklahoma, New Jersey, Arizona, Illinois, and Ohio, to name a few, effectively pressing the brake pedal on data centers. In a six-week stretch early this year, more than 300 data center bills were filed across 30-odd states (Multistate: February 2026), and most importantly, the tone had flipped from courtship to constraint.

Then Gallup asked the question for the first time, in March. Seven in 10 Americans oppose building an AI data center in their area. Forty-eight percent are strongly opposed. Seven percent are strongly in favor.

This is not a localized NIMBY (“Not In My Backyard”) issue. It reflects a national, bipartisan sentiment and serves as the primary barrier to progress.

Why Misconceptions About The Bill Became The Argument

Ask the opponents why, and 20% name costs. Fifteen percent say it plainly: their utility bill.

They have material to work with. Utilities requested a record $31 billion in rate increases in 2025, double the $15 billion sought in 2024, touching 81 million Americans. They asked for another $18.6 billion in the first half of 2026 (Powerlines: January 2026). Residential rates rose by 7.3% between April 2025 and April 2026 and by 18% since January 2025 (Yahoo!Finance: July 2026).

Charles Hua, who runs PowerLines and tracks those filings, has the line that will outlive this news cycle: “Electricity is the new eggs.”

He’s right, and the political consequences have already arrived. Abigail Spanberger and Mikie Sherrill both put utility bills at the center of their 2025 campaigns, and both won. Virginia residential prices had climbed about 13% year over year going into that vote (Politico: November 2026). National reporters who spent a decade treating rate cases as filler are now writing about state commissions ahead of the midterms.

The Uncomfortable Part

This is the critical point where the industry should focus its attention.

Data centers are likely not the culprit.

EPRI looked at the roughly 39% national price increase over the past five years and found that new data centers were not the primary cause. Aging infrastructure due for replacement did much of it. Wildfire and hurricane recovery did more. And natural gas prices rose 56% in 2025, driven substantially by LNG export demand. Gas supplies about 43% of U.S. electricity and sets the marginal price in most hours, so that flows straight through to the bill. One study of the 2015-2024 period found data centers modestly lowered average U.S. retail rates. In some regions with spare capacity, they still do. “[Data centers] are very sophisticated, large customers,” explains Matt Ross, Senior Director of Solutions Engineering at GridX, “and if you give them the right incentives, they can actually be an asset to flatten the peak,” ultimately allowing utilities to do more with what they already have. “If you run more energy through [the same assets], you’re able to recover more revenue. If you’re getting more revenue out of the same infrastructure, that means you have the opportunity to decrease rates while covering the same costs and meeting your return on equity requirements.”

So the sector has a defensible case. It has peer-reviewed backup. It has the game plan to leverage data centers to bend its cost curves. It has the numbers.

Yet, the sector continues to face significant opposition.

Because the 71% of Americans telling Gallup they don’t want one nearby are not conducting a cost-causation analysis. They are looking at a bill that went up, then at a construction site with a substation going in next to it, and drawing a line between the two. That line is intuitive, immediate, and roughly impossible to erase with a fact sheet.

This misdiagnosis deserves attention.

Within the industry, professionals often view affordability as a communications problem. They focus on improving messaging and clarifying what drives costs. This shows they need to do better, and from the interviews we’ve covered in prior issues, it appears the bill might be the best place to drive this communication.

The Exit That Isn’t One

The industry’s other answer is to go around the whole thing. Build generation on site, behind the meter, and skip the queue entirely. About 50 GW of behind-the-meter projects were announced in 2025, and 30% of all planned data center capacity includes plans to build behind-the-meter generation (NewsData: February 2026)

As of mid-2026, only about 2 GW are operational across four projects, with 1,500 MW attributed to xAI’s gas turbines near Memphis. Announcements do not equate to completed projects (Cleanview: 2026).

Furthermore, behind-the-meter generation does not resolve the issue. It increases system-wide bills through fuel market impacts and shared costs that the bypassing entity or center would no longer support. It also fails to address the underlying cause of opposition, as seen in Texas and Memphis. Building isolated solutions does not remove the sector from broader political and economic challenges.

What To Do With This On Monday

For everyone with a job in this business, the practical translation runs something like this.

If you develop capital plans: Include a forecast for public tolerance. Document and quantify it, then stress test this forecast as rigorously as you would a load or fuel curve.

If you handle rate cases: Recognize that cost causation is both a legal principle and a political risk. It is now possible to win regulatory approval but lose public support, as demonstrated in Texas.

If you do siting or community relations: You are no longer downstream of the engineering. You’re the constraint. Budget and staff accordingly, and get there before the opposition group forms, because 833 of them exist and they compare notes.

If you oversee crews: Remember they are trusted messengers within their communities but often lack information. Address this communication gap.

If you serve on a commission: Note that disconnection data and rate request totals are now under public scrutiny and will soon influence political campaigns.