A new data center opposition poll shows how far public sentiment has shifted. In the ABC News/Washington Post/Ipsos poll released on October 21, 2026, 66% of U.S. adults said they oppose building new data centers in the U.S., and 72% opposed one in their own community. For investors, the useful question is not whether the headline is alarming but how that sentiment reaches project timelines, power costs and company guidance. As of October 2026, this guide walks through what the poll measured, what tracked project delays show, why electricity sits at the center of the fight, and how exposure differs across AI infrastructure stocks. It does not forecast any share price.

What the data center opposition poll found

Ipsos fielded the survey on its probability-based KnowledgePanel from October 9 to 13, 2026, among 2,545 U.S. adults, with a margin of error of plus or minus 2 percentage points. According to the Ipsos methodology note, invitations went to 3,861 panelists, interviews were offered in English and Spanish, and results were weighted to Census benchmarks. Several questions were asked of a half sample of 1,271 to 1,274 adults, where the margin widens to 2.8 points. Subgroup numbers carry larger margins still.

The striking detail is that opposition is not a partisan story. Even among Republicans, 58% opposed a data center in their own area, and 80% of adults under 30 did too, along with 69% of those 65 and older. The only group showing majority support nationally was MAGA Republicans, at 53%. ABC noted that data centers have become a midterm election issue while President Trump strongly backs their construction, so federal support and local resistance are pulling in opposite directions.

Separate views on AI from views on buildings

Saying AI is bad for society is not the same as refusing to use AI. In the same poll, Americans were split almost evenly on whether AI is good (50%) or bad (49%) for them personally. The data center numbers are better read as resistance to the local costs of the physical buildout, such as power bills, water use, noise and land use, rather than as a sign that demand for AI services is collapsing.

How local opposition reaches AI infrastructure investing

A poll does not show up in an income statement. It matters when it turns into zoning denials, state legislation, utility rate decisions or lawsuits. Data Center Watch, which tracks local fights over large projects, provides the clearest evidence that this transmission is already happening.

• Q1 2026 report: at least 75 projects worth roughly $130 billion were blocked or delayed amid local opposition, roughly matching all of 2025 in three months. More than 300 state data center bills were filed in the first six weeks of 2026, statewide moratorium proposals appeared in 14 states, and Maine came within one House vote of a statewide ban.
• Q2 2026 report: at least 45 projects worth nearly $68 billion were blocked or delayed. Thirty statehouses introduced legislation, resolutions or executive actions on siting, electricity, water and cost-sharing, and some communities adopted moratoriums before any developer had applied.
• 2023 to Q1 2025 report: $18 billion of projects were blocked and $46 billion delayed, $64 billion in total. Common complaints were higher utility bills, water consumption, noise, property values and green space.

Read these figures carefully. Data Center Watch itself warns that attributing delays only to local opposition oversimplifies the picture, because grid access, regulatory compliance, incentives and market conditions also play a role. Its definitions matter too: a blocked project was halted by a permit denial or withdrawal, while a delayed project is still pursuing approval. Projects often come back in revised form. Tract withdrew a $14 billion Arizona plan in May 2024 and announced another project in the Buckeye area that August, according to the same report.

Why electricity is the flashpoint

Utility bills top the list of complaints because the load is large. A Lawrence Berkeley National Laboratory report released by the U.S. Department of Energy estimated that data centers consumed about 4.4% of U.S. electricity in 2023, or 176 TWh, and could reach 6.7% to 12% by 2028, or 325 to 580 TWh. Serving that demand means new generation and transmission, and state regulators then decide how the cost is shared between residential customers and large loads. DOE said it is working with stakeholders on new rate structures intended to keep power affordable.

Exposure differs across data center stocks

The phrase data center stocks covers businesses with very different sensitivities. The table below maps where local opposition can bite. Actual exposure for any company depends on its own disclosures.

The same headline can cut both ways. Owners who already hold land with grid connections may find those sites more valuable as new approvals get harder, while firms dependent on fresh greenfield entitlements may face cost overruns and delayed revenue. Scale also matters. Data Center Watch listed the $24.7 billion PW Digital Gateway in Prince William County, Virginia, led by QTS and Compass Datacenters, as delayed by litigation as of March 2025. Whether a setback like that is material depends on how much of a company's plan sits in that one location.

Capex updates usually arrive with quarterly results, so it helps to track reporting dates for the largest spenders on the theddari earnings calendar and confirm them against company investor relations pages.

ページの参照

株式theddari.com›株式決算カレンダー企業の決算発表日程とコンセンサス

A checklist for reading AI infrastructure headlines

1. 1

   Read the poll, not just the story

   Check the field dates, sample size, margin of error and exact wording. Do not mix the national question with the in-your-community question, and note when a number comes from a half sample.

2. 2

   Look for a real policy change

   Sentiment becomes a business variable through decisions: a denied rezoning, an enacted moratorium, or a new large-load tariff. Separate bills that were introduced from laws that passed, and record where and when they apply.

3. 3

   Compare with company disclosures

   Search earnings calls and risk factors for permitting, interconnection and community opposition. If capex guidance changes, read whether management blames demand or supply constraints.

4. 4

   Keep macro events in view

   Rate decisions and inflation data move AI infrastructure shares too. Checking the economic calendar keeps you from pinning every price move on a single opinion poll.

ページの参照

株式theddari.com›株式経済指標カレンダー韓国・海外の主要経済指標の発表日程

To see how market chatter around AI infrastructure is evolving, the theddari Telegram stock trends page shows which topics are gaining attention. Popularity is not verification, so return to primary sources before relying on any number.

ページの参照

トレンドtheddari.com›トレンドTelegram · 株式株式チャンネルのリアルタイム投稿フィード

Limits of the poll and the case against overreacting

Neither extreme reading holds up well. A poll is a snapshot, and question wording shapes the answer. Data Center Watch's earlier report cited a survey in which most Americans supported data center construction in the abstract but were skeptical of projects near them, so results are hard to compare across pollsters. Many delayed projects are redesigned or relocated rather than abandoned, and the federal government remains openly supportive of the buildout.

On the other side, the risk can be underrated. Adding the two quarterly counts gives at least 120 projects and roughly $198 billion blocked or delayed in the first half of 2026 (a simple sum of 75 projects and $130 billion in Q1 with 45 projects and $68 billion in Q2; some projects may appear in both). That is not a handful of local exceptions. What matters for an investor is where opposition turns into binding decisions, and how much of a given company's pipeline sits in those places.

In short, the October 2026 poll confirms that the U.S. data center buildout is now exposed to local and state politics in a way it was not a few years ago. Verify the poll wording and margins, follow the quarterly delay counts and state legislation, and match both against each company's capex and permitting disclosures. That sequence lets you judge real exposure instead of reacting to the headline.