QUESTION
We are a mortgage lender in Northern Virginia. An AI Data Center was built in our area, and it is causing havoc with our business. We have branches in Texas, Ohio, and Georgia. Data centers are being built or are under construction in those areas. These are our key states for originating mortgages. Property values have gone down because of these data centers, and that has caused our loan originations to plummet, both in refinances and purchase money.
As far as I can tell, you are the only one telling it like it is in the AI compliance world. My question is indirectly about compliance because we are concerned about our investor partners. Some of them are now pushing back on appraisals and LTV ratios. Quality control is being impacted. I handle sales. Our loan officers are suffering an unprecedented downturn. Lenders that are not affected by AI Data Centers near their markets will eventually feel the same downturn that we are – there's no escaping it!
In our sales meeting, we put together three questions for you.
Which public issues are impacting AI Data Centers?
A few of us believe their expansion is limited by physics.
Is that so?
And how long will this last? Is an AI bubble forming?
COMPLIANCE SOLUTIONS
AI
POLICY PROGRAM FOR MORTGAGE BANKING™
Policies and procedures for
artificial intelligence related to mortgage banking compliance administration,
mortgage loan origination compliance, mortgage servicing compliance, and
industry best practices—including, but not limited to, applicable federal and
state banking and consumer lending laws.
AI FOR MORTGAGE LOAN ORIGINATION - SALES MANUAL
The Sales Manual provides 9
chapters covering the full AI-powered origination funnel in a 90-day
implementation roadmap, with clear team roles and fair lending, FCRA, and
advertising compliance guidance built into every chapter. Sample AI prompts, an
ROI tracking framework, and a compliance glossary.
RESPONSE
I will answer
your questions in three parts, as follows:
· Part I: Public Backlash
· Part II: Physical Limits
· Part III: AI Bubble
In my previous series, The Day of AI Reckoning, I discussed how the public is rejecting the infrastructure and the overall data center opposition. I discussed several compliance risks, some of the compliance ramifications, and why there might be an AI implosion, not a correction.
I get the AI sales pitch: America is building the largest energy-consuming industry of the decade in about five years, and three problems are colliding at once. You have asked three fundamental questions, which I will endeavor to answer. Neighbors are revolting over bills, water, noise, and air. Physics and the grid cap how fast power can arrive and heat can leave. And the bill – well over half a trillion dollars a year – is increasingly being paid with borrowed money.
The three are linked: the physical constraints drive the local backlash, and both shape whether the financing ends in a productive build-out or a bust.
Public issues: who pays, who breathes, who drinks!
The core public complaint is that AI data centers concentrate their costs locally on power bills, air, water, and quiet, while their benefits flow elsewhere. By 2026, that complaint has become one of the few genuinely bipartisan political forces in the United States. Let's break it down categorically, with the caveat that my view reflects current and potential future conditions.
Electricity bills
Data centers are not the only reason power is getting more expensive, but they are now a leading one. I used PJM's independent market monitor, which concluded that expected data center demand drove about $23 billion in customer price increases that will persist through at least 2028. Fortune published an article showing how data centers have hiked electricity prices.
I pushed this conclusion into PolitiFact and found that, for the 2025–26 delivery year alone, it estimated data center demand raised capacity costs by $9.3 billion, or 174%, versus a no-data-center scenario. That correlated with another Fortune article that reported utilities requested a record $31 billion in rate increases in 2025, and residential electricity prices rose 7% that year.
In megawatt terms, when a 300-megawatt campus connects, the grid must add generation, transmission, and substations, and utilities typically spread those costs across every ratepayer. The counter-argument by industry groups and the Energy Secretary – which I think is more political than scientific – is that large loads can lower average rates by spreading fixed grid costs over more kilowatt-hours. I call this the "scapegoat" theory of data center defense, as exemplified by organizations like NetChoice.
Both can be true: it depends on whether new supply keeps pace with new demand and how state regulators allocate costs. The New York Times reported that the Lawrence Berkeley National Laboratory also points to aging grids, equipment costs, and clean-energy mandates as contributors.
Air and Noise
Skipping to a multi-year perspective, developers increasingly build their own gas generation. That is a fact! The Environmental Integrity Project counts at least 74 gas plants planned to serve data centers that ran dozens of unpermitted turbines at its Memphis Colossus site before receiving permits, and has repeated the pattern in Mississippi (see the Senate EPW minority report). xAI, Elon Musk's company, was called out "for its continued operation of unpermitted gas turbines in Southaven, Mississippi, and its pattern of illegal behavior in the region." In Loudoun County, Virginia, residents described a constant low rumble and high-pitched whine from on-site turbines.
Water
Evaporative cooling is enormously water-hungry: a large facility can use up to five million gallons a day, roughly the residential demand of 50,000 people. Indeed, in local permitting records, water is raised more often than electricity! Unfortunately, although operators are shifting to closed-loop and air cooling, the shift saves water but uses more power.
Process and Fairness
As if the foregoing problems are not enough, data center projects are often negotiated under nondisclosure agreements, with neighbors learning late. Tax incentives – such as tax exemptions on servers – buy relatively few permanent jobs once construction ends. Arizona Governor Katie Hobbs has called her state's exemption a $38 million corporate handout.
Political Fallout
Gallup found about 70% of Americans oppose a new AI data center in their area. In Virginia, your home office state, acceptance of local data centers fell from 69% in 2023 to 35% in 2026. Local opposition blocked or delayed roughly 45 projects worth $68 billion in Q2 2026 alone. In response to the widespread opposition, the White House belatedly issued a Ratepayer Protection Pledge in March 2026, and the House passed the Ratepayer Protection Act, a bill "to prevent data centers and other large computational loads from shifting their grid and power generation costs onto existing customers." However, most states have already been working on legislation to make large loads bear their own grid costs. While the states seem proactive, the federal government seems reactive.
_______________
This article, AI Data Centers - Part I: Public Backlash, published on September 30, 2026, is authored by Jonathan Foxx, PhD, MBA, the Chairman & Managing Director of Lenders Compliance Group, founded in 2006, the first and only full-service mortgage risk management firm in the United States, specializing exclusively in residential mortgage compliance.