Data centers > homes?

As we’ve previously noted, Arizona has become a major attractor for data centers and semiconductor manufacturers for a stack of reasons: low humidity, few earthquakes, affordable energy, cheap land, a friendly regulatory environment — and tax breaks.

Data centers pay property taxes in Arizona, but the state exempts certain data-center equipment purchases from state/local sales and use taxes.

Recently, lawmakers seem to be finding bipartisan agreement that it’s time to sunset those deals.

Even one of Hobbs’ gubernatorial challengers, Republican U.S. Rep. Andy Biggs, has joined the bandwagon, supporting Republican Rep. Neal Carter’s bill to end new applications for the tax-break program at the end of 2026.

Data centers don’t provide many local jobs after construction is complete, but they’re still attractive because of property tax contributions. Statewide, an estimated $863 million was collected from data centers in the 2023 tax year — about 1.3% of the state’s annual budget.

The City of Tucson estimated that Project Blue would have contributed $9.7 million in annual city taxes, plus another $152 million in other state and local tax revenues over ten years. That’s a lot of money for elected officials to say “no” to, especially in a state that wants better public services but resists tax hikes.

Plenty of data-center dislike has nothing to do with local economic consideration. More powerfully, data centers have become a symbol of the ultra-wealthy tech elite: Jeff Bezos, Elon Musk, Microsoft, Apple, Google, and more recently, OpenAI. If you don’t love watching these empires expand, you probably love it less when they expand into your backyard.

They’re also widely seen as ugly, soulless architecture — almost like alien presences, taking up a lot of space while never feeling tangibly connected to the community they land in.

In an October report from the Washington Post, experts argued that data centers aren’t the main driver of rising energy costs. Much of the increase comes from higher material prices for transmission and distribution infrastructure. In Western states, those costs are worsened by wildfire damage.

On the contrary, in places with more load growth, rates can fall or stabilize because fixed repair and upgrade costs get spread across more users. But the same experts warn that, over time, enough demand from data centers could move the needle — especially if new generation and new grid upgrades are required.

Last April, we published an edition exploring the history, present and future of Arizona data centers and semiconductor fabrication plants, and their associated water demands.

The core argument: Whatever else we can say about data centers, water usage is often not the crucial variable.

This has borne out in recent months with Project Blue developer Beale Infrastructure announcing they will be moving to a closed-loop air cooling technology which “uses minimal amounts of water that are continuously recirculated.”

It’s a significant improvement which will reduce the majority of the data center’s projected water demands, so long as its energy deal with TEP doesn’t result in an increase in their total energy demands — energy production has its own water footprint.

Even if direct water use becomes negligible, Arizona’s limited groundwater supplies, looming Colorado River cutbacks, and population growth mean the state has to make careful decisions about every new large water and energy user that comes knocking. If a project doesn’t fit into a forward-looking water management strategy, Arizona shouldn’t be quick to welcome it.

But our decisions to welcome or reject projects will have consequences beyond our own backyards.

Growth projections estimate a doubling or tripling in demand for computational power and data storage over the next several years. That means data centers will keep getting built — if not closer to metropolitan areas in the Southwest, then in more vulnerable areas: rural America and countries to our south, as we’re already seeing in places like Mexico, Brazil and Chile.

This is the standard neoliberal option: maintain our quality of life by exporting our costs to poorer countries.

And then there’s the more sci-fi option: put the data centers into Earth’s orbit.

Bezos et al. predict that, by 2035, all new data centers will be floating in the great beyond, where there’s cheap 24-hour solar energy directly from the sun, and no need for cooling — eliminating the two major costs of data center developments.

Whether that sounds like dystopian or utopian sci-fi depends on your comfort with giving the wealthiest people on Earth that much out-of-reach power.

Either way, it’s fair to ask: Is it better to keep these monstrosities somewhere we can keep an eye on them?

Now for the controversial provocation: Maybe it’s a good idea for data centers to use water in Arizona.

Consider Arizona’s new ag-to-urban program, allowing agricultural land owners to sell their water rights to other users, mainly housing developments.

There’s a long-run problem. As water policy expert Kathy Ferris puts it: “Homes cannot be fallowed. Homes are permanent.”

Arizona has long been consuming more water than it has a renewable budget for, and the majority of that demand comes from agriculture. If that agricultural demand simply becomes new residential demand, we’ll create a problem that we can’t solve — a budget that we can’t balance.

Instead, the state needs pathways that permanently reduce water use.

If we require data centers to:

then Arizona could benefit from:

Or we can:

I’ve probably oversimplified and misdiagnosed the situation in multiple ways. But that’s the thing about water: It’s extremely important, complex, and difficult to solve for. If anyone tells you they’ve got it all figured out, you can be sure they don’t.

Maybe there’s low-hanging fruit we can agree on: enforceable limits on industrial water use, and retirement of some agricultural use. Beyond that, Arizona can’t afford to play the game of righteous confidence that has become standard in political contests.

History shows how many present-day problems were created by thinking one step ahead instead of five, and focusing on local impacts instead of downstream impacts. Future Arizonans will inherit the consequences of today’s decisions. The quality of those decisions rests on humility, carefulness and the willingness to sit with uncertainty without rushing into readymade conclusions.

Now I’m starting to sound like Mr. Rogers, so let’s move on to other water news.