
Voters in local initiatives overwhelmingly approved restrictions or pauses on data centers, while Democratic gubernatorial winners who pledged limits include Gina Hinojosa in Texas, Rob Sand in Iowa and Keisha Lance Bottoms in Georgia. JPMorgan's Michael Cembalest says the question after the midterm elections isn't whether data-center creation faces risk, but how much. The report identifies 2028 as the real test for the artificial-intelligence build-out.
The conflict puts decisions over data-center growth in front of voters and elected officials, while JPMorgan assesses what those decisions may mean for the build-out. The source gives no figures for the approved initiatives, no details about the restrictions, and no account of how any pledge will be carried out. It does, however, name both local voter action and gubernatorial winners who promised limits. Those are the concrete points of resistance in a story otherwise told through an investment report.
Who Gets to Set the Limits
Cembalest's central question is about the scale of the risk after the midterms. That framing treats restrictions as a risk to data-center creation, rather than describing what local voters sought or why they approved limits. The report's account offers no explanation of voters' motivations, and no details about the initiatives' language. The distinction matters: the source records a collective decision by voters, then measures its significance against the prospects for an AI build-out.
The three named Democratic gubernatorial winners—Gina Hinojosa, Rob Sand and Keisha Lance Bottoms—pledged to restrict data centers. The source doesn't say whether those pledges became policy, what form any restrictions might take, or how state officials would enforce them. A campaign pledge is not the same thing as a completed restriction. The reported local initiatives, by contrast, were approved by voters, though the article supplies no further details about their reach or implementation.
The Build-Out Meets Local Resistance
The local votes are the clearest example in the source of people acting through a community-level process to limit data-center creation. Their initiatives overwhelmingly passed. The article doesn't identify organizers, describe horizontal organizing, or report mutual-aid efforts, so none can be claimed here. What it does establish is that voters in various local initiatives chose restrictions or pauses, putting a brake on the assumed smooth expansion of data centers.
The source also gives no account of who would bear the costs of construction, or what benefits proponents promised. There are no figures about investment, jobs, or effects on nearby communities. That absence leaves a stark imbalance in the available detail: the report names the financial and political risks to the AI build-out, while offering little about the people whose votes challenged it. It would be speculation to fill in those missing facts.
2028: The Test JPMorgan Sees
Cembalest identifies 2028 as the real test for the artificial-intelligence build-out. JPMorgan's report doesn't spell out what will happen in that year or which restrictions might matter most. It marks a future test, not a settled outcome. The build-out's exposure, in the report's telling, depends on how much data-center creation faces risk after the midterms.
Cembalest also gave a warning on chip stocks this summer. The source doesn't provide the warning's wording or further detail, so its significance can't be expanded beyond that fact. The article names no other corporate or institutional helpers, and it reports no NGO funding or services. Here, the contest is between a build-out assessed by JPMorgan and voters who approved limits, with 2028 left as the year when the report says the stakes will be tested.
- jpmorgan
- michael cembalest
- data centers
- artificial intelligence
- 2028
- data center restrictions
- gina hinojosa
- chip stocks