← The Papers

A policy proposal · Pinal County · May 2026

Building Data Centers
That Pay Their Way

The Utility-Stabilized Data Center Model — protecting residents, the aquifer, and the grid in a desert county, through conditions the County already has the power to attach to a permit.

The idea in one sentence A high-intensity compute campus should cover the residential electric and water cost its presence creates, replace more water than it draws, steward the basin it operates in, and leave the county measurably better than it found it — all through conditions the County already attaches to a permit, at no cost to the County and on legal ground built to hold.

The pipeline is not hypothetical

In roughly six months the Board reshaped thousands of acres of rural and residential land for energy and compute. These decisions will outlast every official who votes on them. This is the demand the framework answers.

0
Acres in La Osa alone
0
Compute at La Osa — 10× a whole electrical district's peak
0
Reported La Osa investment
0
Acre-foot groundwater shortfall ADWR projects this century

La Osa Energy Center

PZ-003-26 / PZ-PD-003-26
3,385 acres
Up to 3 GW compute · up to 2,000 MW on-site gas · WAPA 115kV + APS 230kV · ED4 · Global Water
Board hearing May 27, 2026

Griffin Energy

PZ-PA-014-25
2,685 acres
550 MW solar / 550 MW BESS / 500 MW gas · supports Maricopa DC campus
Rezone pending

Energy Gen & Tech — W Holdings

PZ-PA-015-25
495 of 2,495 req.
Replaces ≈10,600 planned homes · ED3 territory
Rezone pending

Project Midway

PZ-PA-013-25
215 acres
ED3 territory · brings its own power (no firm contract at approval) · ~225 permanent jobs
CPA approved Nov 19, 2025

SRP Marigold Energy Center

SRP board approval targeted Summer 2026
TBD
600 MW solar / 400 MW BESS / up to 675 MW gas · near Stanfield
Pending

Project Bella

Comprehensive Plan approved & rezoned
352 acres
400–440 MW BESS / 480–520 MW gas · high-voltage grid
Approved

Frontier East

Comprehensive Plan approved
160 acres
50 MW solar / 300 MW BESS · near City of Maricopa
Approved

Statuses reflect the public record as of writing. The May 27, 2026 Board outcome on La Osa should be confirmed against the official minutes before circulation.

Two physical constraints

Both point to the same conclusion: commitments expressed as adjectives are worthless. Protection has to be metered, funded, and written into the permit.

5.1 · Water

A closed basin

Pinal sits inside the Pinal Active Management Area. In 2021 ADWR projected a shortfall of more than 8 million acre-feet over the following century — visible already in failing wells, subsidence, and earth fissures — and stopped approving new groundwater-based Assured Water Supply applications. Farmland demand is flexible; farmers fallow in a drought. A hyperscale campus is rigid: once built it locks in a permanent baseline of pumping and removes the flexibility water managers rely on.

5.2 · The grid

A multiple of the whole system

Electrical District No. 3 peaked near 280 MW in August 2025. A 3 GW campus is more than ten times that entire district's peak. La Osa sits in ED4, which generates no power and must buy wholesale and resell. If ED4 finances substations for a campus later downscaled or cancelled, its small residential ratepayer base absorbs the unrecovered cost. A small district should not be the financial backstop for a multi-billion-dollar private campus.

Conditions of entry, not rates

The County can't order a utility to change a bill and can't manage groundwater rights — the Corporation Commission, the utility districts, and ADWR do those things. What it indisputably controls is the permit.

La Osa's rezoning already carries 33 PAD stipulations; conditions of entry are the County's ordinary tool. This paper argues only that the condition set should include protection for residents and the basin, and a standard of stewardship, alongside the usual setbacks and dedications.

The through-line: high-intensity compute does not get to operate here on terms that offload its cost and consequence onto residents and the aquifer. Enforced through conditions the County already controls — not through rates or resources it does not.

Why the County can't just take an ownership stake

The intuitive answer to long-term risk — have the County take equity — is foreclosed in Arizona. Article 9 §7, the Gift Clause, bars a county from becoming a joint owner or shareholder in a private corporation, and from subsidizing one. Under the Wistuber test, sharpened by Turken, Schires v. Carlat (2021), and Gilmore v. Gallego, a public expenditure must serve a public purpose and not be grossly disproportionate to what the public receives — and Schires held that anticipated jobs, growth, and future tax revenue do not count as consideration at all.

Why this proposal is clean: the offsets and obligations are private money flowing to public benefit — the County spends nothing. The reversion vests only in the future, so there is no present joint ownership. The clause polices public giveaways; here, value flows to the public.

The framework, at a glance

Two halves. While a campus operates, it carries obligations in two pillars. When it leaves, it answers a three-layer exit. Every item is a condition of entry.

While it operates — two pillars
Pillar One

Protections

Proportioned to the harm. Cover residential electric & water cost, sized to the campus's own load and locked against rate inflation; replace more water than consumed; pre-fund interconnection & mitigation escrows.

Pillar Two

Stewardship

The standard for operating in this basin. On-site solar whose arrays double as stormwater catchment; required, metered closed-loop cooling — no evaporative waste; reclaimed before potable.

When it leaves — three layers

Don't leave early

A pre-funded penalty scaling inversely to time served.

Don't leave a mess

Durable infrastructure reverts to the public, backed by a decommissioning escrow.

Leave it better

Work off the impact of having existed, through public improvements anywhere in the county.

One rule governs them all: load, not labels

Every obligation that scales — the electric offset, the water replacement — is measured by the campus's actual load and consumption, not by what the project is called or how it's parcelled.

The obligation follows the load. One campus or three, one building called one or three — the kilowatts and acre-feet add up the same, and that total is the obligation.
Aggregated across common ownership and control — regardless of how many LLCs file or how the parcels are drawn.
No merging down, no slicing down. Separately permitted campuses each carry a full obligation; integrated development under common control is treated as one campus so it can't be sliced into "phases."
Measured on permitted capacity, fixed at signing, so it can't be under-reported by running partial loads.

A note on the numbers. Every specific figure — the offset floor, the water and power caps, the committed term, the qualifying load threshold — is illustrative and marked "[Board to set]." What's proposed is the structure, not the settings. The architecture is the contribution; the dials belong to the Board.

Pillar One — Protections

Each protection is proportioned to the harm it answers, so the nexus and proportionality an exaction requires are structural, not asserted.

8.1 · The residential electric offset

The campus covers the residential electric cost in its affected service areas — funded entirely by the operator as a condition of the permit, so the County spends nothing and the Gift Clause doesn't bite. It runs as an operator-funded pool paying the utility or crediting accounts.

Sized to load, with a floor. The offset equals the greater of a fixed minimum [Board to set] or the campus's total load — grid and self-generated alike — in residential-equivalents. The floor guarantees a real benefit from day one; counting self-generation means a campus can't go off-grid to escape the obligation.

Rate stabilization — the lock. The covered residential rate is fixed or capped, so a utility can't inflate a bill a deep-pocketed operator is paying. Cost internalization and rate stabilization work only as a pair. Bounded to the customer base existing at permit issuance, so ordinary growth can't inflate it.

8.2 · Water — cover the cost, and replace more than you take

Water gets a second obligation electricity doesn't need, because water in a closed basin isn't expandable. Covering a bill doesn't put water back in the ground.

The resource side — the teeth. For every acre-foot the campus consumes, it must secure and dedicate at least an equivalent acre-foot back to the basin — by retiring and fallowing grandfathered groundwater rights in the same sub-basin, funding recharge, or bringing in imported or reclaimed water — for a verified net-negative draw, demonstrated before the permit issues rather than promised after the vote. If a campus physically can't secure enough offsetting water in a basin this stressed, that is the basin saying the campus is too large for this place.

Tone on water, plainly: We understand you have to take some water, and we're reasonable about that. In exchange you run the efficient design the industry is already adopting, you prove it with meters, and you put back more than you draw. Take some, waste none, return more.

8.3 · Two place-bound escrows

A Utility Stabilization Escrow, funded before any building permit issues, holds the full capital cost of the interconnections, sub-transmission, and pipeline extensions the load requires — so ED4 recovers its costs if the project stalls and residents are never exposed. An Environmental Escrow takes a recurring, load-proportional fee restricted to aquifer recharge and subsidence mitigation within the Pinal AMA. Both are tied to the land under A.R.S. § 11-1101 and survive any sale or bankruptcy.

8.4 · Resource caps that run with the land

Volume triggers in the PAD overlay bind every future owner. A water trigger pauses later-phase permits automatically if combined draw exceeds a hard cap [Board to set]; a power trigger bars continuous operation without a firm (not interruptible) gas contract, requiring the campus to scale down or draw from on-site storage rather than dumping uncommitted load onto ED4's grid at peak.

Pillar Two — Stewardship

Not offsets and not proportioned to harm — the baseline conduct of operating in a water-stressed desert basin. We ask residents and cities to conserve, harvest, and not waste; a facility that consumes more than a town is not exempt.

☀ Solar = catchment

Meaningful on-site solar whose arrays double as the stormwater catchment surface — one install generates power, captures rain, shades ground, cuts evaporation.

◍ Closed-loop cooling

Required, metered, reported annually against design — turning "near-zero" from a press-release adjective into an enforceable permit condition.

♺ Reclaimed before potable

Where reclaimed water or treated effluent can serve the load, it must — before any potable or fresh groundwater is drawn.

The exit — three layers

During operation the campus pays its way. The exit is where today's risk inverts — from the public inheriting a derelict, scarred site to the public coming out ahead. Each layer is secured by money posted up front.

1

Don't leave early

Abandoning a community promised a campus and a recurring benefit is a breach, not a routine decommissioning — and it should hurt. An early-exit penalty scales inversely to time served — brutal in the early years, tapering toward the end of the committed term — and is pre-funded in escrow so it's real money, not a lawsuit.

2

Don't leave a mess — reversion by asset type

The operator owns and runs everything throughout; the public interest vests only at exit, so there's no present co-ownership the Gift Clause would forbid. What reverts is routed by asset type — a county should hold public facilities, not run a merchant power plant.

3

Leave it better — the impact-deficit ledger

Even a perfectly run campus wore on the place for the length of its life. Net-neutral isn't whole. Existing here runs up an impact deficit the operator works off through public improvement — and not necessarily on the parcel it sits on. A park, paved roads in the county seat, recharge basins, habitat — local benefit residents actually use, over abstract remote offset.

AssetReverts toWhy
Powered, cooled, fiber-connected building shellCountyReusable public facility; servers leave with the operator, the hardened shell does not
Water reclamation plant, recharge facilities, pipelinesCountyDirect public-utility and environmental value
Public-safety facility, dedicated land, roadsCountyStandard public infrastructure dedications
Generation, transmission, substations, interconnectionED4ED4 is the utility in this lane; a county should not run a power plant
Obsolete or hazardous componentsRemovedCleared and the site restored using the decommissioning escrow
10.4 · Security that becomes contribution

The money posted up front need not sit idle for a decade against a betrayal that will probably never come. As the operator clears milestones and the early-exit risk burns down, the capital securing against abandonment converts — milestone by milestone — into credits against the public-improvement obligations it already owes. Protection becomes pre-paid contribution: the security walks down the risk ladder and turns into roads, parks, and recharge as trust is earned. The County is never under-protected; to the operator the capital is never lost — it was parked, and now pays a bill that was coming due. A converted dollar is credited once, never twice.

Why everyone can say yes

The design lets opposite constituencies each find their reason to support it, without those reasons fighting. A protection strong enough to matter usually is strong enough to kill a deal — this decouples them, so the projects get built and the protections are real.

The candidate

Runs on water

Bills covered, every gallon replaced, the county left better — and it's real. Runs hard on resident protection.

The supervisor

Cuts the ribbon

Runs on investment, because nothing here turns a project away. Pro-growth and protected at once.

The developer

Prices it day one

The offsets are priceable up front and the security is parked capital, not lost capital. Permitting certainty is an asset.

The resident

Cost fear answered

The bill is covered, the aquifer is protected, and every recharge basin and paved road is a genuine good.

Traditional incentives vs. the USDC model

DimensionTraditional incentivesUSDC model
Resident experienceHigher bills; opposition each cycleElectric & water cost covered by the operator
The aquiferRigid new draw on a closed basinNet-negative — more water returned than taken
What the County givesTax abatement / public subsidyNothing — it attaches conditions to a permit
Exit / abandonmentStranded cost; derelict, scarred sitePenalty, reversion, county left better than baseline
Legal durabilityVulnerable to Gift Clause challengePrivate funds to public benefit; no county giveaway
Political outcomeRecurring fights, stalled permitsEveryone can say yes; durable across elections

What the Board can do now

None of this requires new state legislation. The County's authority over zoning, PAD overlays, and § 11-1101 agreements is sufficient to adopt these conditions immediately.

Take some — we're reasonable about that. Waste none. Put back more than you took. And leave the place better than you found it.

Enforced through the permit — not through rates or resources the County does not control.

Read the whole thing

The full proposal walks the landscape, the legal ground (Gift Clause, Schires, A.R.S. § 11-1101), every pillar and exit layer, the risks and mitigations, and a full source list for independent verification.

The narrated version

The proposal as a short narrated walkthrough — the problem, the framework, and why everyone can say yes. Narrated by the author.