Quick Answer: A data center backlash is local opposition to building the facilities that power AI and cloud services. Meta president Dina Powell McCormick says the Democratic midterm wins will make new projects harder to build. She argues the opposition is really about how people feel about AI overall, and that projects only succeed when communities see a clear benefit.
Most AI coverage focuses on model releases and chip orders. The less glamorous story is that the buildings housing all that compute have become a political problem. At the WSJ Tech conference, Meta's president said the midterm results will make it "a lot harder" for companies to build data centers. She also admitted the industry "has not done as good a job as we needed to explain the benefits." If you work on infrastructure, land use, or AI deployment, that's a warning worth reading closely.
What Meta Actually Said
As TechCrunch reported, Powell McCormick tied the data center fight to a much larger issue. Data centers were on the ballot in many races, and candidates ran against them. But she framed the buildings as a stand-in for something bigger.
Her core line was simple: "The visible sort of manifestation of AI is a data center, and it's something that people can say, 'I don't want this in my community.'" A server farm is a concrete object you can drive past. AI is abstract, and abstract things are harder to defend at a zoning hearing.
Meta's pitch is that it can win back public trust by showing what the technology does, not just what it consumes. That's a reasonable strategy. It's also the company's own framing, and it's worth testing against what local communities actually experience.
Why the Backlash Is About More Than Buildings
Here's the part that matters most for anyone planning a project: opposition to data centers rarely has one cause. Based on the pattern Powell McCormick described, residents tend to be reacting to several concerns at once:
- Electricity demand and whether household bills will rise
- Water use for cooling and the strain on local supply
- Land use, noise, and the visual footprint of large facilities
- A sense that the local community gets the costs while a distant company gets the profit
- Broader unease about AI itself, including jobs and who controls the technology
The last item is the one that makes the others harder to solve. You can negotiate a water agreement. You can't negotiate away someone's feelings about a technology they didn't vote for.
This is where the lazy advice fails. The common playbook says to offer tax breaks and a few community perks, then the approvals follow. That approach treats opposition as a pricing problem. When the objection is about AI in general, a better deal on property taxes doesn't address the underlying worry. Powell McCormick's point, that the fight is "an underlying bigger issue," suggests the usual incentive package is necessary but not sufficient.
The Deal Structure Meta Says Works
Powell McCormick was specific about what makes a project viable in her view. She said communities "only want data centers when it benefits them in some way." Meta's approach, as she described it, includes several elements:
- Paying for its own electricity so the facility doesn't shift grid costs onto local ratepayers.
- Directing tax surpluses from the project toward teachers and first responders, rather than general budgets.
- Keeping the labor force local so construction and operations jobs go to residents.
- Securing bipartisan support before the project reaches a final vote.
Each of these answers a specific objection. Paying for electricity addresses bill anxiety. Targeting tax revenue to schools and emergency services gives residents a visible return. Local hiring answers the "who benefits" question. Bipartisan backing matters because a project that only one party supports becomes a campaign issue, which is exactly what happened this cycle.
Here's the catch, though: none of these commitments is a guarantee. Electricity arrangements depend on utility regulators, and tax outcomes depend on local budgeting and state law. A company can keep every promise and still see a community turn against it if sentiment shifts. That's the risk Powell McCormick is describing when she says the process will come "under more review."
Scale Raises the Stakes
The numbers help explain why Meta is worried. Powell McCormick said the company operates data centers in 23 states and has five planned projects of 1 gigawatt or more, one of them in Texas. For context, a single gigawatt is a very large amount of continuous power, roughly the output of a major power plant.
Projects at that scale don't slip through quietly. They require new transmission, long utility negotiations, and often state-level approvals. Each one creates a new set of local stakeholders who can organize, testify, and vote. A single 1 GW campus can become a statewide story in a way that a smaller facility never would.
This next part matters more than it looks. Scale cuts both ways. A large project brings bigger tax surpluses and more jobs, but it also concentrates every concern: power draw, water, noise, and traffic. If you're planning a campus this size, you need the community case built years before the first permit application, not after the first public hearing.
Is the Infrastructure Still Worth Building?
Meta's bet on AI depends heavily on its consumer AI agent, Muse. Powell McCormick said Muse currently has 10 million users, and the company wants it to scale far beyond that, eventually serving billions. Whether that happens is an open question, and it's worth being honest about that uncertainty.
Meta's fallback is more interesting. Powell McCormick argued that even if consumer AI falls short, the compute itself keeps its value. In her words: "We're getting calls every day to lease that compute, and so one way or another, we're seeing a path to monetization."
That's a real hedge. Compute capacity can be leased to other companies, which means a data center isn't a dead asset if one product underperforms. The trade-off is that the economics shift from owning a consumer business to running infrastructure, a lower-margin, more utility-like model that still depends on steady power and public permission to operate. Meta's own statement that the approach must make the local case is a reminder that the hedge still needs a license to build.
Frequently Asked Questions
what is a data center backlash?
A data center backlash is organized local opposition to building or expanding data centers. Residents, officials, and candidates object to power demand, water use, land use, and costs to local utilities. In 2026, Meta's president said such opposition was a major issue in the midterm elections and could make future projects harder to approve.
how do you win community approval for a data center?
Start with the benefits residents can see and verify. Pay for the electricity your facility uses, direct tax revenue to visible public services, hire locally, and seek bipartisan support before the vote. Meta says these steps separate successful projects from failed ones, though no combination guarantees approval.
why do people oppose data centers in their area?
Most objections combine practical and broader concerns. Residents worry about higher power bills, water strain, noise, and the sense that a distant company gets the profit while the community carries the costs. For many people, opposition also reflects unease about AI as a technology, which a single project can't resolve.
can data centers pay for their own electricity?
Yes, companies can commit to paying the full cost of the power their facilities draw, which helps protect household ratepayers. Meta says it pays for its own electricity in successful projects. The actual arrangement depends on utility rules and state regulators, so the commitment needs to be written into the deal to carry weight.
Data center backlash isn't going away after one election cycle. If you're planning AI infrastructure, the lesson is to earn local permission early, with benefits people can see and verify, and to treat the public's view of AI as part of the project's risk. Pick one project in your pipeline and write down the three concerns its neighbors are most likely to raise, then answer each in plain language before you file anything. Share this with anyone wrestling with a zoning fight right now.