Rejected 7-0 by the city over water and secrecy, then rebuilt one jurisdiction over. The most useful part is not why it died, but exactly how it came back.
Updated 2026-08-14 · maintained as new filings and rulings land
Location
near Tucson, Pima County, AZ
Developer
Beale Infrastructure (Blue Owl)
End user
AWS (reported, unconfirmed)
Size
600 MW · ~$3.6B · 290 ac
Water (reported)
up to ~283M gal/yr
Status
Revived via county, Dec 2025
Before the autopsy
A $3.6B campus got rejected 7-0 by the city. Best guess: what sank the vote?
The setup
Project Blue was a hyperscale campus near Tucson, developed by Beale Infrastructure and backed by Blue Owl Capital, with Amazon Web Services reported as the end user, though the sponsors never confirmed it. The plan: roughly 290 acres, up to ten buildings, about 2 million square feet, 600 MW, and around $3.6 billion in investment, phased from 2026 to 2029.
The flashpoint was water. The original design used evaporative cooling that needed City of Tucson water, and the developer sought to have the county land annexed into the city to get it. Reporting put consumption at up to roughly 283 million gallons a year, framed against thousands of households, in an arid region already anxious about Colorado River cuts. Layered on top: secrecy. The end user was withheld and the project advanced under non-disclosure agreements, which turned uncertainty into suspicion.
Timeline
Jun 2025Pima County approves a land deal, but it is contingent on the city annexing the site and providing water service.
Jul 2025Public scrutiny escalates around water use and the undisclosed end user.
Aug 6, 2025The Tucson City Council votes 7-0 to reject the project. Because the June county deal depended on annexation and water, the rejection effectively kills it.
Aug 8, 2025Beale cancels its Tucson reclaimed-water system expansion plans.
Late Aug 2025Beale files an energy supply agreement with the Arizona Corporation Commission and pivots to building on unincorporated Pima County land, which needs no city annexation.
Dec 16, 2025The Pima County Board of Supervisors approves a binding Memorandum of Agreement with Beale on a 3-2 vote. It includes independent annual verification that Beale matches 100% of the project's energy with renewables. Cooling is switched to a closed-loop system that does not draw city water.
Dec 24, 2025The developers close on the purchase of the 290-acre county land. The campus rejected 7-0 in August is alive again by Christmas.
Cause of the rejection, in detail
Three vulnerabilities compounded. First, a water-dependent design in a drought region gave opponents the single most resonant grievance available. Second, secrecy about the end user converted normal uncertainty into active distrust; a large, unnamed corporate water user is a story that writes itself. Third, approval ran through a discretionary city vote: annexation was a gate the council could simply close, and it had both the political will and the legal authority to do so, unanimously.
The four mistakes that lost the city
Picked a water fight it did not need. Designing around scarce municipal water in a drought region created the most inflammable objection possible, before a single benefit was on the table.
Let secrecy define the narrative. Withholding the end user did not lower the temperature; it handed opponents a "what are they hiding" frame.
Chose a jurisdiction that could veto. Routing the deal through a politically exposed city council created a single, winnable target for the opposition.
Led with the ask, not the concessions. No water-neutrality plan or community-benefit story arrived before the opposition organized. By the time commitments came, distrust was already set.
The comeback, which is the real lesson
What makes Project Blue worth studying is that the developer did not simply appeal or relitigate. It re-engineered around each fatal point.
Removed the dependency. Switching to closed-loop cooling neutralized the "corporate water grab" argument at the root.
Changed the venue. Moving to unincorporated Pima County removed the city's veto and reset the politics.
Bought political cover. A binding agreement with independent 100% renewable-energy-match verification gave county supervisors something defensible to vote for.
Locked power early. Filing the energy supply agreement with the Arizona Corporation Commission removed the energy-uncertainty attack.
Signals that were public beforehand
A water-dependent design in a metro already fighting over supply.
An undisclosed end user, which communities reliably read as something to hide.
An approval path that required annexation and a council vote, a discretionary gate rather than a right.
The insight to carry away
A resource dependency you do not control is a kill switch. Water, city power, or annexation. Design it out (closed-loop, on-site generation, by-right land) rather than spending political capital defending it.
Venue selection is strategy, not paperwork. A discretionary vote is a veto point. A by-right or friendlier jurisdiction removes the single target opponents need.
Secrecy raises opposition; disclosure plus binding commitments lowers it. Verifiable community benefits buy the political cover a deal actually needs.
A rejection is not always a death. A re-engineered project in a friendlier venue can survive, but only if the fatal dependency is genuinely removed, not just re-argued.
Take it to your IC
We keep a board-ready one-pager of this autopsy, including the comeback playbook: the timeline, the deal-level detail, and every source, formatted to drop straight into an investment-committee memo. Want it sent over?
On its way. Check your inbox.
Updates
2026-08-14 Autopsy published: full timeline, cause of rejection, the comeback playbook, and sources.
This file is maintained. As new filings, rulings, or deal detail surface, they get logged here.