This project is not anti-data-center.
It is against dark government — public decisions made out of public view.
It is against the absence of transparency — public records that exist but are not made accessible, public processes that are shortcut, public officials who act on the public's behalf without informing the public.
It is against undue corporate influence in matters that affect citizens — when the public interest is set aside because a private interest has more political weight, more lawyers, or more time.
It is against the anti-democratic activities of elected and monied interests — the routine accumulation of small bypasses, fast-tracked permits, sealed agreements, executive sessions, threats of litigation against local governments that try to listen to their constituents.
Data centers happen to fall under all of this right now in Texas because the political and regulatory machinery around them has tilted, in real and documentable ways, against the people who live near them. That is the narrow case the published deck makes.
The broader case is that the same patterns repeat — solar farms, pipelines, gravel pits, freight rail, water transfers, cell towers, eminent-domain takings. What this project is for is citizens having the same access to the data and the analysis as the people on the other side of the table have always had.
If a 2,075-acre industrial assembly with a 75-megawatt fast-tracked air permit, no public notice, and an implied gigawatt-scale buildout is brought to a community after the fact — that is the dark-government problem, not the data-center problem.
If a public meeting confirms that a state senator threatens litigation against a county for considering a moratorium — that is the anti-democratic-activities problem, not the developer problem.
If a corporate group operates fourteen sites across eleven Texas counties under entity names designed to obscure the common operator — that is the transparency problem, not the gas-turbine problem.
A community that wants a data center, with eyes open, should be free to have one. A community that does not, should be free not to have one. What is not acceptable is for the choice to be made for them, in silence, by people they did not elect.
The data assembly, statistical models, cartographic rendering, and much of the writing supporting this project were produced with the assistance of Anthropic's Claude. This is disclosed openly, both here and on the landing page of the publication.
There is a category of objection that says: it is hypocritical to use AI to research a story about the consequences of AI buildout. I disagree with that framing, and I want to be explicit about why.
A hammer is not a house. A spreadsheet is not a forensic accounting. A neural network is not a research project. In each case the tool is necessary, sometimes uniquely capable, and never sufficient. What matters is whether the human directing the tool has the discipline and the judgment to use it for something the public should want to exist.
That means: parsing 18 PDFs of tax notices, joining 254 county polygons to 101 groundwater-district boundaries, querying public databases by their documented APIs, running a cost model whose formulas are traceable in published source code, drafting narrative passages that are then verified against the cited sources line by line.
Every numerical claim in the deck is reproducible from source code in the repository. A human reviewed every output before publication. Where AI assisted in writing, the writing is verifiable against primary sources cited in the same paragraph.
Every quote from a public official traces to a published meeting record or a public document. Every numerical claim traces to a public-records source listed in the methodology.
The same critique would have applied to typewriters in 1880, to copy machines in 1960, to spreadsheets in 1985, and to the World Wide Web in 1995. Tools change what an individual citizen can do on their own. Some uses of any new tool are abusive, fraudulent, or lazy. That is not an argument against the tool. It is an argument for paying attention to the use.
If anything, the asymmetry of who has access to powerful analytical tools is itself a public-interest concern. Large corporations have had teams of analysts, lawyers, and lobbyists for as long as those professions have existed. An individual citizen working in their living room with a public internet connection and an AI assistant can, today, produce a documented public-interest analysis that would have been impossible for a private citizen to assemble five years ago. The democratic implication of that is substantial, and on balance favorable.
The skill is in the user. The discipline is in the methodology. The accountability is in publishing everything — code, data, methodology, and sources — so that anyone can check the work.
Democracy is not a set of institutions. It is a set of practices. Among those practices, one of the most important and most fragile is the routine exercise, by ordinary citizens, of the right to read the public record and draw their own conclusions from it.
That right is not in any immediate danger. But it is being constricted in ways that are easy to miss: through processes that are technically public but practically inaccessible, through permits that are technically subject to comment but published without notice, through agreements that are technically signed by elected officials but negotiated by parties the public never sees.
The remedy is patient, sourced, documented work, made available to whoever wants to look. That is what this project is. That is the standard it tries to hold itself to. And that is the standard it would invite anyone challenging its findings to meet in return.