Here’s what’s on our plate today:

• 🧪 The year AI became a voting issue.

• 📰 The AI psychosis debate, AI's new visual frontier, AI's weirdest product yet.

• 💬 Prompt of the Day: Map how AI could swing competitive 2026 districts.

• 🗳️ Poll: What turns out the angriest AI voters?

Let’s dive in. No floaties needed…

The Laboratory

TL;DR

• Anxiety crossed party lines: 71% of Americans think AI is moving too fast, with Democrats and Republicans nearly even in their concern. The old ideological split around AI has collapsed under shared economic pressure.

• Power bills are doing the politicking: U.S. electricity costs are up 40% since 2021, utilities requested $30B in rate increases in 2025 alone, and data centers are absorbing enormous grid capacity while ratepayers absorb the tab.

• Communities are blocking the buildout: 48 data center projects representing $156B in investment were stalled by local opposition in 2025. More than 188 active opposition groups now span 17 states, and the coalition is genuinely cross-partisan.

• Young workers are the angriest bloc: Only 18% of 14- to 29-year-olds feel hopeful about AI. Early-career workers in AI-exposed roles have experienced a 13% relative decline in employment since 2022.

• Pennsylvania is the test case: Three toss-up House districts are ground zero for incumbents caught between a $20B Amazon expansion and constituents watching jobs and utility costs shift simultaneously.

The year AI became a voting issue

When hundreds of residents packed a county commission meeting in Tremonton to oppose a large AI data center backed by Kevin O’Leary, they were not protesting artificial intelligence in the abstract. Their concerns were immediate and local: pressure on the town’s water supply, strain on the electricity grid, and the possibility that the character of their community would be permanently altered by infrastructure built to serve distant technology companies. The commissioners approved the project anyway, but residents immediately organized a ballot referendum for November.

That sequence of events, public opposition followed by approval and then political mobilization, is increasingly becoming commonplace across the United States. And driving them is the transition of the AI industry from a technological curiosity to something capable of impacting voters across political leanings.

For most of the past three years, public concern about AI moved through familiar channels. Academics warned about bias and automation, policymakers drafted frameworks, tech CEOs testified before Senate committees, and went home to keep building. However, in 2026, the debate has turned material, with AI’s infrastructure expansion showing up in monthly electricity bills, and its effects on entry-level hiring arriving fast enough for a generation of workers to feel them before they have had a chance to adapt.

The old partisan frame is gone

According to an Economist/YouGov poll conducted in May 2026, 71% of Americans believe AI is advancing too quickly, including 77% of Democrats and 68% of Republicans. The numbers are notable because they reveal how evenly anxiety now cuts across partisan lines. Three years ago, the concerns were more clearly divided: Democrats tended to worry about safety and societal harm, while Republicans focused more on government overreach and regulation. But now that distinction is eroding fast.

YouGov’s own tracking data shows negative views of AI rising from 34% to just over 50% over the same period, suggesting that skepticism is no longer tied to a single ideology or concern. And increasingly, the unease appears broader, driven by the sense that technology is moving faster than the public can adapt to it.

Among young people, anxieties about AI run especially high, as they enter the workforce at a time when automation is increasingly threatening their prospects.

According to a Gallup survey of 1,572 people ages 14 to 29, conducted in February and March 2026, only 18% feel hopeful about AI, down from 27% the prior year. Anger rose from 22% to 31% over the same period.

Gallup senior education researcher Zach Hrynowski offers one explanation for the shift: college seniors and young professionals who spent years, and often substantial amounts of money, preparing for careers are now watching those same career paths begin to restructure before they have had a chance to enter them properly. For many younger workers, the anxiety around AI is no longer theoretical or ideological. It feels immediate, economic, and increasingly personal, with the sense of displacement building so quickly that it could become a political grievance heading into November.

The electricity bill is on the ballot

According to Fortune, power bills in the United States have risen 40% since 2021, with utilities requesting more than $30B in rate increases in 2025 alone, affecting 81M Americans. Data centers, which require enormous amounts of electricity to run and cool the servers that power AI systems, account for a significant and growing share of those increases, and their concentrated, localized demand has made them the most visible focal point of public anger over energy costs. A single hyperscale facility, the kind that runs large AI models, can draw between 100 and 1,000 megawatts of electricity, roughly equivalent to the power needs of 80k to 800k homes.

The cost structure compounds the anger in a particular way. When a utility builds new infrastructure to serve a data center, that investment is typically spread across all customers in the service area, not just the operators. Tax revenue from a new facility may accrue only to the host county, while everyone on the same grid bears its power costs. In at least one region, the Mid-Atlantic, there is direct evidence that data center demand contributed to rate increases, and the pattern is expected to spread.

Morgan Stanley analysts, writing about political risks ahead of the midterms, called public pushback on this dynamic a “binding constraint” on buildout, and that language reflects a supply chain problem rather than a reputational one.

According to Fortune, capital expenditure projections for U.S. hyperscalers, namely, Microsoft, Amazon, Google, Meta, Oracle, and CoreWeave, sit at $785B for 2026, with nearly $1T projected for 2027. The industry has never bet bigger, and the communities where that bet must land have veto power.

When the ground pushes back

The clearest form of pushback from communities has been protests, which start quietly at first and then grow into a scale that is difficult to ignore. According to 10a Labs’ Data Center Watch initiative, as reported by Fortune, at least 48 data center projects representing $156B in investment were blocked or stalled by local opposition in 2025 alone. In Q1 2026, more than 20 additional projects were canceled at a record quarterly pace, bringing the total to more than 188 opposition groups now active across 17 states.

What makes this movement politically durable, rather than a series of local fights that fade after a zoning vote, is that it is genuinely cross-partisan. Conservative landowners in rural Indiana, progressive activists in Tucson, and suburban homeowners in Maryland are arriving at the same position through entirely different reasoning, and that breadth makes the coalition difficult for either party to absorb or dismiss.

The extent of the protest can be gauged by examining the legislative response to the growing problem.

Maine’s legislature, in April 2026, passed the nation’s first statewide moratorium on data centers above 20 megawatts, pausing approvals until November 2027. And though Governor Janet Mills, a Democrat, vetoed it over the absence of a carve-out for a local project, at least 12 other states have filed similar bills.

On March 25, 2026, according to a press release from his office, Senator Bernie Sanders and Representative Alexandria Ocasio-Cortez introduced the AI Data Center Moratorium Act in Congress, a federal freeze on all new construction until legislation covering worker protections, environmental impact, and civil liberties is signed into law. The bill is unlikely to advance in the current Congress, and it has not been designed primarily as legislation. However, it can still serve as the foundation for a political coalition built around a shared set of demands that future candidates will be forced to address.

Where the votes are

For the clearest picture of how AI could be impacting politics, keep an eye on Pennsylvania, where the tension between the voters and the legislators on the topic becomes concrete.

The state already hosts more than 100 data centers, and Amazon has committed $20B to significantly expand that footprint, with the active support of Democratic Governor Josh Shapiro. The expansion is concentrated in the eastern part of the state, where Republicans hold four competitive House districts—the Cook Political Report rates three of them toss-ups.

According to CNBC, Chris Borick, director of the Muhlenberg College Institute of Public Opinion, said bipartisan backlash around data centers has “layered another challenge” onto incumbents already navigating difficult terrain. That leaves Republicans caught between an administration that wants them to support aggressive AI buildout and constituents watching their utility bills climb.

The job displacement question adds a third layer, and it is where the anger connects back to the Gallup data on young people. According to the Stanford HAI AI Index 2026, early-career workers in the most AI-exposed occupations have experienced a 13% relative decline in employment since the widespread adoption of generative AI in 2022, even as overall employment continued to grow. That figure points to something structural in entry-level hiring, and the people it affects most are exactly the cohort showing up in polling as the angriest: workers who did everything they were told to do and are finding the rules changed mid-game, which is the kind of story that travels at the ballot box.

The industry’s bet, and its limits

For its part, the AI industry has countered the current protests by stating that data centers create local construction jobs, generate significant tax revenue, and underpin services millions of people use every day.

According to the Associated Press, the Data Center Coalition, an industry group, has warned that a moratorium would raise costs for American families and eliminate hundreds of thousands of high-wage jobs.

These arguments are not going away, and it is clear they will be well-funded: dueling super PACs are reportedly preparing to spend hundreds of millions of dollars making opposing cases to voters ahead of November.

However, what is less clear is whether those arguments have traction at the zip code level, where the debate is actually being decided.

The Trump administration’s preferred mechanism, a voluntary pledge signed in March 2026 by Amazon, Google, Meta, Microsoft, OpenAI, Oracle, and xAI, commits the companies to fund their own power generation rather than pass grid costs to ratepayers. The pledge carries no penalties for noncompliance and no auditing requirements, and the communities that have spent two years in zoning hearings and public comment sessions have become reasonably good at distinguishing a binding commitment from a press release.

Pennsylvania’s competitive districts in November will be the first real signal. If incumbents in those seats lose partly on data center and job-displacement grievances, the calculation for every elected official who has publicly backed AI buildout changes, along with the industry’s model for predicting where community resistance ends, and community acceptance begins. Whether that happens, and how quickly the industry adjusts if it does, is the question the rest of 2026 will answer.

Tuesday Poll

Prompt Of The Day

Bite-Sized Brains

• The AI psychosis debate: Researchers are split over whether "AI psychosis" is a real clinical phenomenon or a moral panic, as more cases of chatbot-fueled delusion reach the press.

• AI's new visual frontier: A new BBC short explores how AI is reshaping the way we create and consume visual media, blurring the line between real footage and synthetic imagery.

• AI's weirdest product yet: An AI-powered crypto cannabis vape has arrived, a near-perfect snapshot of every overhyped tech trend colliding into one absurd device.

The Toolkit

• Deepgram: Speech-to-text API built for scale, handling real-time transcription and voice intelligence for production apps.

• Descript: AI-powered audio and video editor that lets you edit recordings by editing the transcript like a doc.

• Drift.ai: AI add-in for Excel that makes financial models context-aware and queryable, built for investment banking, VC, and corporate finance teams.

Rate This Edition