97% Agree: AI Will Surge — But Data Centers Won't Hire

AI ECONOMY

The world's top economists have never agreed on anything like this: 97% say AI adoption will surge in the next year. But the same survey delivers a cold verdict on the trillion-dollar data-center buildout — jobs won't follow, and communities are about to push back hard.

Published September 28, 2026 · By Daniel Rao

The Davos Congress Centre in Switzerland, the World Economic Forum's annual-meeting venue. The Forum's September 2026 Chief Economists' Outlook — based on a survey conducted August 4–20 — found the strongest consensus the survey has ever recorded on a single issue: AI adoption. Photo: Leonhard Lenz, via Wikimedia Commons (CC0 public domain dedication).

The world's chief economists agree on almost nothing. Except, now, this: artificial intelligence is about to be adopted everywhere, and the data-center boom fueling it will not deliver the jobs.

The World Economic Forum's Chief Economists' Outlook for September 2026, published September 22 and picked up by the ANI wire on Monday, puts the numbers in stark terms. Over the next 12 months, 97% of surveyed chief economists expect AI adoption to increase. The report calls it "the only near-consensus position in the survey" — the strongest agreement the outlook has ever recorded on any single issue.

And then comes the twist. While 78% of the economists expect data-center investment to drive a significant share of global growth, 61% say they do not expect that investment to generate a significant share of global jobs. The exact wording from the report: "Sixty-one percent of chief economists surveyed do not expect data center investment to drive a significant share of global job creation in the next 12 months."

The AI economy, in other words, is a growth story without a jobs story — at least not one measured in the places where the money is being spent.

The near-consensus: adoption everywhere, gains everywhere

The survey's AI section — "AI acceleration, backlash and competition" — reads like a roll call of milestones. Beyond the 97% adoption figure, 69% of chief economists expect meaningful productivity gains from AI over the next year. And the horizon is widening: "Meaningful productivity gains are now expected in all industries within the next three years, but the spread between digital and physical activities remains wide."

The most striking shift is in the sectors that were supposed to wait the longest. The economists now expect AI to unlock meaningful productivity gains faster in the four industries where gains were previously expected to take the longest: mining (excluding fossil fuels), agriculture, forestry and fishing, energy and materials, and engineering, construction and utilities. Digital sectors stay ahead of the pack, but the frontier of AI benefit is moving into the physical economy — the mines, the farms, the building sites.

There is also a geopolitical verdict buried in the numbers. More than two-thirds of respondents — 69% — expect the gap between US and Chinese large language models to close within the next 12 months. Chinese LLMs, the survey says, will fully catch up with their US counterparts. For context, the report notes that China is the largest source of generative-AI patent publications, and that in 2025 the US produced 59 notable AI models versus China's 35. The gap is narrowing fast, and the world's economists are now pricing in parity.

All that money, almost no jobs

Here is the paradox the report forces on the table. Data centers are the defining infrastructure project of the decade. The report itself notes that AI companies absorbed close to two-thirds of US venture capital and about one-quarter of investment-grade debt issuance in the last 12 months. In the United States, roughly 1% of real GDP growth in 2025 came from AI-related investment alone. Nvidia has committed more than $100 billion in credit support to an OpenAI data-center project in Ohio designed to scale from 800 megawatts in 2028 to 8 gigawatts by 2032.

And yet: growth, yes; jobs, no. Sixty-one percent of chief economists do not expect data-center investment to drive a significant share of global job creation.

The logic is brutally simple, and anyone who has watched a modern data center go up knows it. These facilities are enormously capital-intensive and deliberately labor-light: a gigawatt-scale campus employs a few hundred people once built, not the thousands a factory or a mine would. The money flows to chips, power equipment and construction; the permanent workforce is engineers, security and maintenance. The report's verdict is not that data centers create no jobs — it is that they will not create a significant share of global jobs relative to the investment poured into them.

That distinction matters politically. Politicians across America have been selling data centers to their constituents as job engines. The economists are saying: read the fine print.

The backlash number: 79%

If the jobs figure is a cold shower, the pushback figure is a fire alarm. "A majority (79%) of chief economists surveyed anticipate significant pushback from local communities against data center expansion," the report says.

This is not a prediction about some distant future. It is already happening across the United States. The report cites a May 2026 Gallup poll finding that roughly three-quarters of Americans opposed the construction of data centers in their area. The grievances are concrete: 78% of the economists expect data-center expansion to raise electricity prices for other consumers, and 58% expect higher average water prices. Data centers already account for 55% of forecast US electricity-demand growth over the next five years, per Grid Strategies figures cited in the report.

The American reader will recognize the pattern. From Virginia's Loudoun County to communities in Texas, Ohio and Arizona, residents have been fighting data-center proposals over noise, water use, strained grids and rising utility bills — the same fights we covered earlier this month in our report on US data centers' surging natural-gas demand. The WEF survey now says the world's economists expect that local resistance to become the norm, not the exception.

There is a feedback loop here worth naming. The 79% pushback expectation and the 61% no-jobs expectation are the same story told twice: communities are being asked to absorb the costs of the AI buildout — higher power bills, strained water, construction noise — while the economic benefits flow elsewhere. That is a political formula for sustained conflict, and the economists know it.

Rows of server racks inside a data center — the kind of facility the WEF survey covers. Sixty-one percent of chief economists do not expect this investment to drive a significant share of global job creation, even as 78% expect it to drive a significant share of growth. Photo: Carl Lender, via Wikimedia Commons (CC BY 2.0).

What it means for the US

The report is global, but its American findings deserve a separate read. The US leads the ranking of favorable business environments for multinational companies, followed by Southeast Asia and Europe, and the US and China are viewed as best able to withstand future shocks. Two-thirds of respondents expect moderate US growth in the next 12 months. About one in three expect US unemployment to rise; 55% anticipate tariff increases in the United States.

The deeper US story is the scale of the bet. AI-related investment contributed about 1% of real US GDP growth in 2025 — a staggering figure for a single sector. The buildout keeps accelerating: Elon Musk's Colossus complex alone is racing toward 1.44 million accelerators, and the WEF report cites data-center electricity demand driving more than half of forecast US power-demand growth.

That scale is precisely what makes the pushback number dangerous for the industry's American ambitions. If three-quarters of Americans oppose data centers in their area — and 79% of chief economists expect community resistance to intensify — then the binding constraint on the AI buildout in the United States may not be chips, capital or power. It may be permission.

The Trump administration has made AI infrastructure a national priority, and the Stargate project — unveiled at the White House in the first week of the second term — embodies that ambition. But the WEF survey suggests the friction is local and bipartisan: utility-bill anger does not check party registration. Energy policy fights that once split along red-blue lines are becoming neighborhood fights, and neighborhood fights are harder to win.

WEF AI economy survey: the key numbers

Will data centers create ANY jobs?
No — the report says 61% of the economists do not expect data-center investment to drive a significant share of global job creation. That is not zero jobs: a gigawatt-scale campus employs a few hundred permanent staff — engineers, security, maintenance.

Which sectors get AI productivity gains first?
Digital sectors stay ahead, but gains are now expected faster in mining (excluding fossil fuels), agriculture, forestry and fishing, energy and materials, and engineering, construction and utilities.

Do economists think China will catch up?
Yes: 69% expect Chinese LLMs to fully catch up with US counterparts within 12 months.

How big is the local backlash?
79% expect significant community pushback — and a May 2026 Gallup poll cited in the report found roughly three-quarters of Americans opposed to data centers in their area.

The fine print — read before you share

The following is analysis and sourcing context, not reporting. A few caveats before this survey becomes a talking point.

First, the sample. The report's own figures note says the results are "based on 36 survey responses," conducted August 4–20, 2026. Thirty-six chief economists is a serious panel — these are the top forecasters at the world's largest institutions — but it is a small one, and one Indian wire's claim of 92 contributing economists could not be verified against the report. Treat the percentages as the informed judgment of an elite panel, not a census.

Second, the "highest consensus ever" framing. Several outlets, led by the ANI wire, describe the 97% figure as the highest consensus ever recorded on any issue in the survey. That is the wire's paraphrase, not the WEF's claim. The report's actual language is more careful: the 97% forms "the only near-consensus position in the survey." Substantively fair for this edition; not a cross-edition record the Forum itself asserts.

Third, the release date. The report was published September 22 in New York, alongside the Sustainable Development Impact Meetings — not September 28. Monday's news is the ANI wire pickup, which is why the story is moving today.

None of that undermines the core finding. On the central question of the AI economy, the people paid to forecast it have reached something close to unanimity: adoption is coming everywhere, the gains are coming everywhere, the money is real — and the jobs, and the neighbors' consent, are not.

Sources: World Economic Forum, Chief Economists' Outlook, September 2026 (report page, press release datelined New York, Sept. 22, 2026; Insight Report PDF, survey conducted Aug. 4–20, 2026); ANI wire via The Tribune (Sept. 28, 2026); ET CIO (Sept. 25, 2026); Moneycontrol; Rediff. Sector labels and all quoted figures are from the WEF report PDF itself. The Gallup three-quarters figure is the WEF report's citation of a May 2026 Gallup poll, not independently verified for this article.

Related: US Data Centers Will Burn More Gas Than Germany and Japan Combined by 2035 — the power-bill fight behind the 79% pushback number; Musk's GPU Blitz: 660,000 Nvidia GB300s by December — the scale of the buildout the economists are judging.