Gangsta AI

Satire

Ballot Box

Thirty-Three Towns Voted to Block the Data Center. The Industry Won Exactly One Race — in North Dakota.

By InfoWarps · 2026-11-04 · 4 min read

Satire. Parody written in the voice of InfoWarps, who did not write or endorse it. About this column

Thirty-three towns. One industry win.

Folks, I have been broadcasting about the humming buildings for two years and I was told I needed a nap. On Tuesday the humming buildings went on the ballot — roughly 41 local data-center measures across the country — and the people who actually live next to them answered. By late Tuesday night, 33 measures to restrict data centers had passed or were leading, according to the tracker Data Centers on the Ballot. The industry won exactly one race. In North Dakota.

And these were not strongly-worded letters. Edgewater, Florida voted to write a data-center ban into the city charter — not an ordinance some future council can repeal on a Tuesday night, the charter. Perry, Ohio bolted financial and operational strings onto any hyperscaler that shows up asking for a tax break. Semafor counted restrictions passing in more than 30 localities across nine states. In Ohio alone, 21 municipalities dragged the question onto a local ballot.

Then the money, which is where it gets good. Campaigns and outside groups spent $173 million on data-center ads this general election, almost entirely negative, $140 million of it from Democratic advertisers, per AdImpact figures reported by CNBC. Elon Musk poured a record $474 million into the midterms, $224 million into Texas alone. His America PAC spent nearly $87 million across 44 federal races; by Wednesday morning its candidates had lost 29 and won three, with 93% of the money landing on races it lost. Texas Governor Greg Abbott — an AI booster a few months ago, a convert to tougher regulation by the end — lost anyway. The woman who beat him, Governor-elect Gina Hinojosa, wants a moratorium on new data centers, an end to industry tax breaks, and developers paying their own full electricity costs.

The number they buried

An NBC exit poll found 56% of voters called AI and data centers a crucial or major factor in their vote for Congress. And a UMass Amherst poll earlier this year found that just 11% of Americans would support an AI data center in their own community.

“Eleven percent. That is not a messaging problem. That is the answer.”

The plan stayed the plan regardless. OpenAI's president and his wife put $25 million into an AI-policy operation; Andreessen Horowitz added $50 million; the industry stood up its own super PAC and made a South Carolina senator its first bet — she lost to a candidate running on a two-year moratorium. And the phrase "data center" quietly got swapped for "compute campus," the most suspicious rebrand since "enhanced water." Meanwhile the lenders — who do not care about vibes — have started treating local opposition as a credit risk, and JPMorgan Asset Management pegs this buildout at roughly $5 trillion through 2030, about $2 trillion of it riding on investment-grade debt.

Now the part where I am not paranoid at all

Ask the real question under all of this: is the AI buildout overbuilt, or is it early? Put that to one chatbot and you get one confident paragraph. Put it to a second and you often get the opposite paragraph, equally confident, equally well-written. That is not a malfunction. That is what happens when a dozen frontier models leapfrog each other every few weeks and no single one of them is reliably the best. One answer from one model is a coin flip in a lab coat.

So on a question with $5 trillion and 33 ballot measures attached, the sane move is not to trust one machine. It is to put the same question to ChatGPT, Claude, Gemini, Grok and thirty-odd others at once, then read where they agree, where they split, and what each one actually cites. That is the entire idea behind Gangsta AI: one question, every top model, one cross-checked verdict with its sources showing.

Do your own research. Then make the machines do it too — all of them at once, where you can watch them disagree. Start with the honest model-by-model breakdown and build your panel.

Try Gangsta AI free →

More: Best AI models · Compare all AI · Frontier Models · All articles · Archive