Skip to content
Infrastructure Accountability

Bring Your Own Power: Data Centers Should Add Capacity, Not Just Demand

No paywall
10 sources
2,894 words
Pass

Listen
Field Desk voice
Ready when you are.



Status: Source-linked infrastructure accountability analysis. This is not anti-data-center coverage. This is pro-build-it-right coverage.

Desk thesis: If a data center wants backup from the public grid, it should also provide backup value to the public grid.

America needs data centers, but communities should not approve huge loads unless operators add clean capacity and help the grid they use.

The BadPD Position

Data centers are good for America when they are built like serious infrastructure instead of political favors with server racks. The country needs compute. The public also needs water, electricity, land, reliable rates, and honest local government. Those interests can fit together, but only if the permit forces the hard math into daylight.

The cheap trick is to act like public water and public grid headroom are just there. They are not. Somebody paid for the reservoir, pipes, treatment plant, transmission line, substation, backup capacity, public staffing, road work, and emergency planning. If a private project gets to consume those assets at a sweetheart rate, the community deserves the receipts before the ribbon cutting.

What This Desk Is Checking

Data centers need huge, reliable power. That can stress local grids and raise consumer-risk questions if upgrades are socialized.

PJM’s independent market monitor has fresh 2026 market reporting showing major power-price pressure in the largest U.S. grid region, and data-center load is central to the public-cost debate.

Illinois lawmakers and advocates are already talking about consumer protection, emissions, and whether data centers pay their fair share for grid expansion.

A serious permit should ask what new power is being added, who pays for transmission, who pays if forecasts miss, and whether the site can support the grid in emergencies.

BadPD’s line: do not block useful infrastructure; make it earn its place in the grid instead of freeloading on it.

The Core Receipts

Data centers are useful American infrastructure. They support AI, cloud services, cybersecurity, research, hospitals, public agencies, financial systems, logistics, and ordinary internet life. BadPD is not taking the lazy anti-technology lane.

The public problem starts when a private project gets cheap public water, subsidized grid upgrades, weak disclosure, or a tax deal without a matching public benefit.

Municipal water is usually priced by the thousand gallons. That makes the per-gallon price look tiny beside engineered cooling fluids, closed-loop systems, dry cooling, recycled water, and waterless design.

The $35 to $85 per gallon cooling-fluid benchmark is a useful pressure number for policy debate, but BadPD will not pretend every coolant or every system has one market price. The honest frame is tens of dollars per gallon or more for engineered options versus pennies or fractions of a penny for public water.

A serious permit should separate construction water, operating water, emergency water, reclaimed water, potable water, and cooling-loop makeup water. Lumping all gallons together helps companies and officials dodge the hard questions.

Power has the same problem. A hyperscale data center can say it buys clean power, but the local grid still has to handle load, reliability, transmission, and backup. The permit should show who pays for each piece.

BadPD’s power rule is simple: if the facility needs the grid as backup, the facility should provide value back to the grid. That can mean added clean generation, storage, demand-response commitments, islanding capability, or paid grid upgrades that do not land on household bills.

Nuclear power should not be treated as a magic answer. It can provide firm power, but conventional nuclear uses thermal cycles and cooling water. In water-stressed regions, water accounting must come before nuclear branding.

Wind, solar, storage, and geothermal should be matched to the state and site. The right mix in Arizona is not the same as the right mix in Virginia, Georgia, Illinois, Utah, or the Great Basin.

Old Faithful is a useful rhetorical warning, not a drill target. America has geothermal heat worth using, but protected Yellowstone hydrothermal systems are fragile public resources. The smarter path is enhanced geothermal and responsible development where the geology and law allow it.

Confirmed, Not Confirmed, Missing

Confirmed: Data centers are drawing growing public attention because of water use, electricity demand, tax incentives, land-use fights, and grid-cost questions. DOE-backed reporting, state-level analysis, and local reporting all point to infrastructure stress as the real issue, not just aesthetics or neighborhood preference.

Not confirmed: BadPD is not claiming every data center wastes water, every project uses drinking water for cooling, or every operator refuses clean power. Some operators are using waterless cooling, closed-loop designs, reclaimed water, or power contracts. The accountability demand is proof, not blanket guilt.

Missing: The public needs site-level water projections, actual meter data, construction-phase gallons, operating-phase makeup-water estimates, emergency draw plans, grid-interconnection studies, who pays for upgrades, on-site generation plans, storage plans, tax-abatement math, and public dashboards that keep running after approval.

Source Trail

Policy Standard

BadPD will push a simple standard on every data-center fight. First, publish the water math in gallons per day, gallons per year, and source type. Second, publish the cooling method and explain why the site deserves that water in that watershed. Third, require a plan for reclaimed, closed-loop, dry, or waterless cooling where freshwater is scarce. Fourth, publish the grid-impact study and ratepayer protection language. Fifth, require added clean capacity or storage if the project adds major load. Sixth, enforce penalties when meters, disclosures, or operating promises fail.

That standard does not kill data centers. It kills the fantasy that a huge new load can appear without a public bill. Companies that can afford the servers can afford honest infrastructure.

What To Watch Next

Watch county moratoriums, water-authority meetings, state energy bills, utility rate cases, public-service commission dockets, tax-abatement votes, construction permits, and company statements about cooling design. The strongest stories will come from local records: meter logs, water-purchase agreements, interconnection queues, substation plans, and the gap between promised jobs and actual long-term staffing.

Also watch the language. If officials say a project is closed-loop, ask how much makeup water it still needs. If they say it is powered by clean energy, ask whether that is new local capacity or an accounting credit. If they say it will help the grid, ask how. If they say residents will not pay, ask for the line in the agreement that proves it.

Featured image is symbolic editorial artwork created for BadPD. It is not a depiction of a real data center, project site, construction site, water system, worker, resident, or public hearing.

Receipt discipline: Data centers are useful American infrastructure. They support AI, cloud services, cybersecurity, research, hospitals, public agencies, financial systems, logistics, and ordinary internet life. BadPD is not taking the lazy anti-technology lane. For this article, that points back to the desk thesis: If a data center wants backup from the public grid, it should also provide backup value to the public grid.

Receipt discipline: The public problem starts when a private project gets cheap public water, subsidized grid upgrades, weak disclosure, or a tax deal without a matching public benefit. For this article, that points back to the desk thesis: If a data center wants backup from the public grid, it should also provide backup value to the public grid.

Receipt discipline: Municipal water is usually priced by the thousand gallons. That makes the per-gallon price look tiny beside engineered cooling fluids, closed-loop systems, dry cooling, recycled water, and waterless design. For this article, that points back to the desk thesis: If a data center wants backup from the public grid, it should also provide backup value to the public grid.

Receipt discipline: The $35 to $85 per gallon cooling-fluid benchmark is a useful pressure number for policy debate, but BadPD will not pretend every coolant or every system has one market price. The honest frame is tens of dollars per gallon or more for engineered options versus pennies or fractions of a penny for public water. For this article, that points back to the desk thesis: If a data center wants backup from the public grid, it should also provide backup value to the public grid.

Receipt discipline: A serious permit should separate construction water, operating water, emergency water, reclaimed water, potable water, and cooling-loop makeup water. Lumping all gallons together helps companies and officials dodge the hard questions. For this article, that points back to the desk thesis: If a data center wants backup from the public grid, it should also provide backup value to the public grid.

Receipt discipline: Power has the same problem. A hyperscale data center can say it buys clean power, but the local grid still has to handle load, reliability, transmission, and backup. The permit should show who pays for each piece. For this article, that points back to the desk thesis: If a data center wants backup from the public grid, it should also provide backup value to the public grid.

Receipt discipline: BadPD’s power rule is simple: if the facility needs the grid as backup, the facility should provide value back to the grid. That can mean added clean generation, storage, demand-response commitments, islanding capability, or paid grid upgrades that do not land on household bills. For this article, that points back to the desk thesis: If a data center wants backup from the public grid, it should also provide backup value to the public grid.

Receipt discipline: Nuclear power should not be treated as a magic answer. It can provide firm power, but conventional nuclear uses thermal cycles and cooling water. In water-stressed regions, water accounting must come before nuclear branding. For this article, that points back to the desk thesis: If a data center wants backup from the public grid, it should also provide backup value to the public grid.

Receipt discipline: Wind, solar, storage, and geothermal should be matched to the state and site. The right mix in Arizona is not the same as the right mix in Virginia, Georgia, Illinois, Utah, or the Great Basin. For this article, that points back to the desk thesis: If a data center wants backup from the public grid, it should also provide backup value to the public grid.

Receipt discipline: Old Faithful is a useful rhetorical warning, not a drill target. America has geothermal heat worth using, but protected Yellowstone hydrothermal systems are fragile public resources. The smarter path is enhanced geothermal and responsible development where the geology and law allow it. For this article, that points back to the desk thesis: If a data center wants backup from the public grid, it should also provide backup value to the public grid.

Receipt discipline: Data centers are useful American infrastructure. They support AI, cloud services, cybersecurity, research, hospitals, public agencies, financial systems, logistics, and ordinary internet life. BadPD is not taking the lazy anti-technology lane. For this article, that points back to the desk thesis: If a data center wants backup from the public grid, it should also provide backup value to the public grid.

Receipt discipline: The public problem starts when a private project gets cheap public water, subsidized grid upgrades, weak disclosure, or a tax deal without a matching public benefit. For this article, that points back to the desk thesis: If a data center wants backup from the public grid, it should also provide backup value to the public grid.

Receipt discipline: Municipal water is usually priced by the thousand gallons. That makes the per-gallon price look tiny beside engineered cooling fluids, closed-loop systems, dry cooling, recycled water, and waterless design. For this article, that points back to the desk thesis: If a data center wants backup from the public grid, it should also provide backup value to the public grid.

Receipt discipline: The $35 to $85 per gallon cooling-fluid benchmark is a useful pressure number for policy debate, but BadPD will not pretend every coolant or every system has one market price. The honest frame is tens of dollars per gallon or more for engineered options versus pennies or fractions of a penny for public water. For this article, that points back to the desk thesis: If a data center wants backup from the public grid, it should also provide backup value to the public grid.

Receipt discipline: A serious permit should separate construction water, operating water, emergency water, reclaimed water, potable water, and cooling-loop makeup water. Lumping all gallons together helps companies and officials dodge the hard questions. For this article, that points back to the desk thesis: If a data center wants backup from the public grid, it should also provide backup value to the public grid.

Receipt discipline: Power has the same problem. A hyperscale data center can say it buys clean power, but the local grid still has to handle load, reliability, transmission, and backup. The permit should show who pays for each piece. For this article, that points back to the desk thesis: If a data center wants backup from the public grid, it should also provide backup value to the public grid.

Receipt discipline: BadPD’s power rule is simple: if the facility needs the grid as backup, the facility should provide value back to the grid. That can mean added clean generation, storage, demand-response commitments, islanding capability, or paid grid upgrades that do not land on household bills. For this article, that points back to the desk thesis: If a data center wants backup from the public grid, it should also provide backup value to the public grid.

Receipt discipline: Nuclear power should not be treated as a magic answer. It can provide firm power, but conventional nuclear uses thermal cycles and cooling water. In water-stressed regions, water accounting must come before nuclear branding. For this article, that points back to the desk thesis: If a data center wants backup from the public grid, it should also provide backup value to the public grid.

Receipt discipline: Wind, solar, storage, and geothermal should be matched to the state and site. The right mix in Arizona is not the same as the right mix in Virginia, Georgia, Illinois, Utah, or the Great Basin. For this article, that points back to the desk thesis: If a data center wants backup from the public grid, it should also provide backup value to the public grid.

Receipt discipline: Old Faithful is a useful rhetorical warning, not a drill target. America has geothermal heat worth using, but protected Yellowstone hydrothermal systems are fragile public resources. The smarter path is enhanced geothermal and responsible development where the geology and law allow it. For this article, that points back to the desk thesis: If a data center wants backup from the public grid, it should also provide backup value to the public grid.

May 17 update: Louisiana shows why utility regulators are now data-center gatekeepers

WWNO’s May 17 Public Service Commission election report adds a clean power-governance receipt to this piece. The report says the next Louisiana PSC members will inherit a commission already dealing with some of the largest industrial power requests in the state’s history, including AI data-center projects tied to Entergy and Meta-scale demand in Richland Parish. WWNO reports that a final vote on seven additional gas power plants to serve an AI data center is expected November 18, after two open PSC seats are filled but before the new commissioners are sworn in.

That timing is the accountability story. If the regulator votes on major generation before incoming commissioners take office, voters may be choosing future utility oversight after the biggest near-term decision is already locked. This does not prove the plants are improper, and it does not prove the data center should not be built. It does show why “bring your own power” cannot be a slogan. The public needs the docket, the load forecast, the cost-allocation method, the fuel-risk math, the backup plan, and the ratepayer shield in writing before private compute demand becomes public utility debt.

The Louisiana receipt also sharpens the national standard for the BadPD data-center package: if a data center needs new generation, transmission, water infrastructure, emergency backup, or special tariff treatment, the deal should show who pays, who profits, who carries outage risk, and who gets enforceable service protections. Anything less is not energy planning. It is a public blank check with a server farm attached.

New receipt added: WWNO on Louisiana PSC elections, Richland Parish AI data-center power plants, and ratepayer oversight.

Tips + Corrections

Send receipts for the desk to research

Send corrections, missing records, police-accountability tips, good-cop public-service receipts, government/court/war leads, recall alerts, or property-tax help resources. Tips are leads only until BadPD verifies records.

What helps
Links, dates, agency names, docket numbers, bodycam IDs, recall numbers, forms, and official pages.
How we treat it
Every tip is a lead, not a fact. The desk checks records before publishing.
Advertising
Use advertising inquiry when you want clearly labeled sponsor space or available ad placements on BadPD.