Saagar Enjeti: data center backlash will go national in 2026 as electricity costs hit rural America

Summary

Data Center Backlash Is Brewing — and Saagar Enjeti Thinks It Goes National by 2026

The political collision between the AI infrastructure build-out and ordinary Americans is no longer a fringe concern. Electricity costs are the flashpoint, and the pressure is already visible at the local level before it reaches Washington.

Virginia, where data centers consume 40% of all electricity, and Oregon, where the figure is around 33%, are the early fault lines. Rural communities near data centers are reporting power bill increases of up to 267%. The issue surfaced in Virginia's most recent gubernatorial debate, with both the Republican and Democratic candidates calling for action. Prince William County residents are overwhelmingly opposed to further development, and Tucson, Arizona is also moving against new projects.

Enjeti's core argument is structural. The U.S. grid is aging, electricity is not a solved problem the way it is in China, and massive capital expenditure on data centers is creating a finite-resource competition — particularly in rural areas — with almost no offsetting local job creation. Construction workers and specialist tradespeople get paid, but communities see little else. The dynamic echoes the Amazon warehouse expansion of the past decade, where the company became the second-largest U.S. employer while simultaneously signaling plans to eliminate or forgo hundreds of thousands of jobs.

The political timeline Enjeti maps is specific. Democrats are expected to win a meaningful number of elections in 2026, and anti-tech sentiment is already a small-d democratic issue surfacing at town halls. On the right, he expects Tucker Carlson to lead the rhetorical charge, with Marjorie Taylor Greene and Josh Hawley following. He flags David Sacks, the White House AI adviser, as a central figure on the pro-AI side — but notes that Sacks's argument that AI is necessary for any GDP growth is politically toxic framing: it translates in practice to "grow the economy so you can lose your job."

Bernie Sanders called for a breakup of OpenAI just before this segment aired, making him the first U.S. politician to do so explicitly. Enjeti reads this as the socialist left and the populist right converging on the same destination through different routes — a horseshoe dynamic where "enough" becomes the shared political message.

AI Valuation Concerns Mirror the Dot-Com Era

Enjeti draws a direct parallel between current tech stock performance and 1999. S&P 500 gains attributed to AI stocks are tracking at similar levels to the dot-com peak, with today's GDP figures running even higher than then. He flags round-tripping vendor finance deals — citing Cisco's collapse as a historical reference point — and expresses concern that Masayoshi Son's deepening involvement with OpenAI has the structural feel of the WeWork dynamic, where external pressure to go bigger inflated commitments beyond what fundamentals could support.

His broader concern is concentration risk. The entire U.S. retirement system, interest rate assumptions, and GDP projections are banking on a single sector delivering transformative returns — a bet he considers genuinely dangerous regardless of whether any individual AI company survives.

The Productivity Reality vs. the AGI Pitch

The gap between what AI is being sold as and what it actually delivers in practice is widening. The real-world use cases Enjeti encounters — auto meeting summaries at Accenture, marginally better research tools, browser-based search improvements — are useful but do not justify valuations premised on civilizational transformation. His framing: AI is delivering a slightly better version of existing internet services, competing for the same screen-time economics as AdWords did in 2002 and 2003, while the marketing language promises something categorically different.

He applies the same critique to the social cost side of the ledger. Jonathan Haidt and the growing institutional push to ban phones from schools represent a mass elite-level revolt against the attention economy. AI is accelerating that backlash rather than escaping it, particularly because the wealthiest technology insiders — who are most aggressively promoting AI adoption — are also the most likely to strictly limit their own children's screen time.

Sports Media as a Secondary Bubble

On the media landscape, Enjeti's sharpest call is that sports media is currently the most inflated independent bubble, driven almost entirely by FanDuel and DraftKings advertising spend, which he estimates at 80 to 90% of revenue for most sports podcasts. He expects a significant wipeout across the category when sports betting advertising pulls back, with only two or three properties — he names Pardon My Take — surviving at scale.