[ THE ULTIMATE GUIDE ]
Every serious argument, for and against, side by side — and every one of them opens. Underneath: the charts, a deal simulator you can run yourself, and the full Alabama file. Start here, though: why we need these things at all.
[ 01 — WHY WE BUILD ]
A data center is a building full of electricity and math. That sounds boring right up until you notice what humanity is currently pointing all that math at. Five reasons the answer to "should we build these?" starts as yes — and then the rest of this page is the argument over on what terms.
Federal interest payments accrued since you opened this page. The Treasury is running about $3 billion a day on interest alone — roughly $1.0 trillion in FY2026, more than the entire defense budget. Nobody is cutting their way out of that number. The only historical exit from a debt that size is growth — and the growth on offer right now runs on compute.
For most of history, finding a drug meant a chemist with a hunch and a decade. Then protein folding — a problem that had eaten 50 years of biology — got solved by a model. AlphaFold turned structure prediction from a PhD thesis into an API call.
It kept going. Isomorphic Labs shipped IsoDDE in February 2026, a unified drug-design engine that roughly doubled AlphaFold 3's accuracy on the hardest cases — 50% versus 23.3% where the target looks nothing like anything in the training data. Those are exactly the targets nobody could drug before. The company now has 17 active programs across oncology, immunology, and cardiovascular disease, with its first AI-designed cancer drug tracking into Phase 1 trials by the end of 2026.
Here's the part that matters for this page: that pipeline is compute-bound, not idea-bound. Optimized training runs for protein-folding models now use 2,080 H100 GPUs to finish in ten hours what originally took a week. Every one of those GPUs lives in a building with a substation and a cooling plant. When you ask whether we should build data centers, you are also asking how fast we get to run the experiment that finds the thing that kills the cancer.
Every era has one industry that decides who sets the rules. Steel did it once. Oil did it for most of a century. Right now it's compute — because compute is what turns into weapons research, drug pipelines, code, logistics, intelligence analysis, and the models that everyone else ends up renting.
America is currently ahead, and the margin is not small. US and allied manufacturing capacity for advanced AI processor dies runs an estimated 35 to 38 times China's, quality-adjusted. Huawei's yields sit around 5–20% against Nvidia's 60–80%. Absent chip exports, US installed compute capacity in 2026 would be more than ten times China's. On spending: America's four largest hyperscalers plan roughly $650 billion this year, while Alibaba — China's most aggressive builder — committed $53 billion across three years.
But a lead in chips only becomes a lead in capability if somebody plugs them in. A GPU in a warehouse is a paperweight. The data center is the part where the advantage becomes real — and it's the part that has to be built here, on our grid, under our law, by people we can subpoena. You cannot offshore a substation.
The numbers, plainly. Federal debt held by the public reaches roughly 101% of GDP in 2026. The FY2026 deficit runs about $1.9 trillion, 5.8% of GDP. Interest alone costs $1.0 trillion this year — which is more than the $885 billion for national defense and more than the $708 billion for Medicaid. CBO has interest reaching $2.1 trillion by 2036.
There are exactly three ways out of a debt-to-GDP ratio like that: raise taxes enough to close a 5.8%-of-GDP gap, cut spending enough to close it, or make the denominator bigger. Every country that has actually escaped a debt load like this did the third one. You grow.
And growth is what this buildout is currently producing. AI-related capital expenditure contributed roughly 1.1 percentage points to GDP growth in the first half of 2025 — outpacing the American consumer as an engine of expansion. By Q1 2026, investment in AI data centers, hardware, and networking amounted to 1.4% of US GDP, up from 0.7%, and roughly half of headline GDP growth traced to categories where the AI buildout is the dominant source of new demand.
That's a real, live, measurable thing pulling in the right direction on the worst number in American public finance. It is not the whole answer. It is the only part of the answer currently showing up in the data.
The drug pipeline gets the headlines, but the machines are general. The same racks that fold proteins also run hurricane track models that give the Gulf Coast an extra day of warning, screen candidate battery chemistries and superconductors without touching a beaker, model plasma confinement for fusion, plan the transmission upgrades the grid has been deferring since the 1990s, and tutor a kid in Cullman who doesn't have a calculus teacher.
None of that is speculative technology. It's existing technology waiting on capacity. The bottleneck between "we could run that experiment" and "we ran it" is almost always a queue for GPUs. Every megawatt that comes online shortens somebody's queue.
This is the difference between a scarcity mindset and a growth mindset, and it's worth being precise about it. Scarcity says the pie is fixed, so the fight is over slices. Growth says we have repeatedly, historically, and recently made the pie bigger — and the mechanism for making it bigger has always been building physical infrastructure that lets people do more with the same day.
The most common mistake in this whole debate is treating "don't build it" as the neutral option — the safe default where nothing changes and nothing is risked. That's not one of the choices on the table.
If a town says no, the campus goes to the next county. If a state says no, it goes to the next state. If America says no, it goes somewhere with worse labor law, dirtier power, and a government that would very much like to read what's on those drives. The compute gets built. The only open question is who gets the jobs, the tax base, the leverage, and the say.
Which is exactly why the rest of this page is not a sales pitch. Saying yes to the technology and yes to any deal offered are completely different positions, and the gap between them is where a community either gets rich or gets used. So: here is every argument, both directions, with the numbers attached.
[ 02 — THE SCALE ]
Both sides of this argument have an incentive to lie about the size of the number. Boosters call it a rounding error; opponents call it the end of the grid. Here is the actual measured curve, from the Department of Energy's own lab, with the projection range drawn as a range — because that's what it is.
US_DATA_CENTER_ELECTRICITY // 2014–2023 measured · 2024–2028 projected range
Share of all US electricity that data centers used in 2023 — 176 TWh. Real, but not yet the grid's main event.
The projected 2028 range. The spread between those two numbers is the entire policy argument — one is manageable, one is a crisis.
Growth from 2014 to 2023, from 58 TWh to 176 TWh. Then the AI server buildout bent the curve again.
Data centers' share of TVA's entire industrial load in 2025, up from 1–5% a few years earlier — and TVA expects it to double again by 2030.
[ 03 — THE LEDGER ]
Forty-six arguments, matched left and right. Every card opens into the actual numbers plus the steelman — the strongest response the other side has to that specific point, written by someone who takes it seriously. Filter by topic, search the text, and mark which arguments actually move you. The scale keeps score.
Mark arguments as MATTERS or DECISIVE as you read and the scale will tip. Nothing weighed yet — the beam is level.
▲ THE CASE FOR
▼ THE CASE AGAINST
[ 04 — THE DEAL SIMULATOR ]
Almost nobody actually disagrees about whether a 300-megawatt building should exist. They disagree about who pays for the power, where the water goes, and what happens if the company walks. So build one. Move the sliders and watch the same facility turn from a windfall into a fleecing and back.
// Estimates from public benchmarks, not a specific project. Assumptions and formulas are in sources & methodology.
[ 05 — THE ALABAMA FILE ]
More than $3.5 billion is already committed to Alabama ground, on a grid that's cheaper than the national average, in a state where — until this year — the incentives were among the most generous in the country and most counties had no legal power to say no. Then the 2026 legislature moved. Click a site.
ALABAMA_SITES // built · building · contested
Here's the thing most of the national coverage missed: two of the three biggest complaints on this page have already been legislated in Alabama — this year.
SIGNED APR 16, 2026 · EFFECTIVE JUN 1, 2026 · CAP BEGINS JAN 1, 2027
Caps new data-processing-center tax abatements at 20 years unless the operator signs a binding community investment agreement, and requires at least 20 new jobs averaging $40,000 a year including benefits. This is the clawback-and-conditions structure critics have been demanding for a decade, and it's now the law.
EFFECTIVE OCT 1, 2026
Sets up Public Service Commission review so that a very large AI data center covers its own incremental power costs. This is the single most important line in the entire ratepayer argument: it puts the cost of new generation on the customer who caused it, instead of socializing it across everyone's power bill.
UNRESOLVED
Alabama counties under 600,000 residents have no zoning authority under the state's Dillon's Rule structure. Homewood, Springville, Leeds, and Cullman enacted municipal moratoriums in June 2026, but a 50 MW facility in Morgan County proceeded anyway — because there was no zoning to violate. A community that cannot say no cannot negotiate.
UNRESOLVED
Nothing in the 2026 package requires closed-loop cooling or public reporting of water withdrawals. The difference is not marginal: a comparable campus runs about 22,000 gallons a day closed-loop versus 5 million evaporative. That's a rule worth writing before the next groundbreaking, not after.
TVA_INDUSTRIAL_LOAD // data centers' share, and where it's headed
// TVA is pursuing 6.2 GW of new generation — the largest capital program in its history — citing data centers and population growth. Alabama residential power runs about 16.77¢/kWh, roughly 9% below the national average of 18.44¢. That gap is the state's leverage. It is also the thing most at risk.
[ 06 — THE HONEST VERDICT ]
If you read all forty-six arguments, a pattern shows up: almost every genuine downside on the right-hand column is a terms problem, not a technology problem. Each one has a known fix that some jurisdiction has already written down. Four terms decide whether a data center is a windfall or a transfer.
New generation and transmission caused by one customer get billed to that customer, through a large-load tariff — not smeared across every residential ratepayer. Alabama's SB270 does this. It removes the single largest legitimate complaint.
Require closed-loop cooling and public water reporting. It's a 200-fold difference in daily consumption, it's commercially available today, and the operators building it voluntarily prove the requirement isn't a dealbreaker.
Every abatement carries clawbacks tied to jobs, wages, and capital, with a hard sunset. Alabama's HB399 now caps at 20 years and demands a binding community agreement. Incentives should be a contract, not a gift.
A community that legally cannot refuse has no negotiating position, and gets the worst version of every other term. Alabama's zoning gap is the biggest unfixed item on this page — and it's the cheapest one to fix.
Get the terms right and this is the best deal a rural county has been offered in fifty years. Get them wrong and it's a substation with a tax bill attached.
Both of those outcomes are available from the exact same building. That's the whole argument. The technology is not the variable — the contract is.
So be relentlessly pro-building and relentlessly pro-terms at the same time. Those have never been opposites, and treating them as opposites is how a state ends up with neither the growth nor the guardrails. Alabama got two of the four right this year. The future can be genuinely, unreasonably good — but only for the people who show up to the negotiation.
On honesty: the projections on this page are projections. The 2028 range is 325 to 580 TWh because the people who measure this for a living genuinely do not know which it will be. Anyone quoting you a single confident number for 2030 — in either direction — is selling something. Weights you set in the ledger are stored in your own browser and go nowhere.
// jeremymsparks.com — last reviewed August 2026