Howdy, from the editor

Today’s edition of the Ground Party Papers is a big one.

Really big.

It covers: local water issues, global power struggles, artificial intelligence, government surveillance, data centers of course, and even factories on the moon.

The purpose is to give you a working sense of the “big picture” around data centers and the many ways they are shaping our lives — locally and far beyond our little patch of desert.

This edition is longer than usual, so please treat it like a road trip. Bring snacks. Stretch your legs. Take breaks to go outside and stare at mesquite trees for a while.

There’s a lot in here, but the practical takeaway is actually pretty simple:

One of the best things we can do in response to this strange new era is build more of the good old-fashioned human stuff. Less time streaming and scrolling. More potlucks, campfires, local food, local friendships, town halls, shindigs — the squishy human activities we’re built for.

But if you’re up for it, strap in for a little tour of our new data-powered world. I can’t promise tidy answers. I’m fairly certain they don’t exist. My job is to look and see what’s floating around out there, and try to report back to you as clearly as I can.

We’ll begin our strange little field trip by diving into the deep end.

Grab your flashlight, thermos, and thinking cap.

Sign up for the only subscriber-funded newsletter in Cochise County.

Chapter 1 : The Technocracy
Chapter 2 : Data & AI News Cycles
Chapter 3 : Inside the Machines
Chapter 4 : Water, Energy & Money
Chapter 5 : A Path to Power

The Technocracy

How many of the people below do you recognize?

One way to understand the modern story of power is to “follow the money.”

Another way is to follow information, technology, military power, or the glowing surveillance rectangle in your pocket.

Either way, you’ll keep finding the same people at the top: the techno-elites.

Unlike 100 years ago, these barons are not operating in smoke-filled backrooms. Instead, they speak freely about their domination plans, in interviews, podcasts, speeches, and investor calls. They even write books about them.

What are the elites planning while the rest of us are busy with jobs, bills, entertainment, culture wars and homesteadmaxxing?

“Disruption.”

Walmart disrupted local groceries. Uber disrupted taxis. Cars disrupted horses and carriages. Guns disrupted swords. Drones disrupted guns.

People protest the disruptions. They keep happening anyway.

The richest men in the world got rich by disrupting private industries.

Now, they’re talking about disrupting governance.

And because public trust in government and major institutions is so low, those ideas can sound less shocking than they might have decades ago.

So, what are they saying?

—

Elon Musk

Elon Musk’s grandfather was a leader in the Technocracy political movement, which argued that engineers and technical experts should hold the highest political power, because modern civilization runs on the technical systems they built and understand.

Technocrats say that politicians are better at winning elections than designing the machinery of governance. They can stir up voters with propaganda about the “other side”, but they can’t actually build systems that work.

Most Americans already believe the government is untrustworthy.

Musk and other tech elites are offering a different answer: rebuild the operating system.

Musk wants X (formerly Twitter) to become an “everything app” — a single online hub for messaging, media, banking, and much of daily digital life. He wants people to use X to manage their “entire financial life” and has predicted it could become “half of the global financial system.” (Watch the interview.)

This was part of Musk’s original vision for X.com, the online bank he founded before it merged into PayPal. After being pushed out as CEO, Musk went on to build Tesla, SpaceX, Starlink, The Boring Company, Neuralink, xAI, and Terafab.

That is quite a résumé.

Along the way, Musk has become the richest and most powerful private actor on Earth. He owns more than half the satellites in orbit, builds underground tunnels, funds political campaigns, wins government contracts, develops brain-computer interfaces, and is trying to make millions of humanoid “Optimus” robots.

With DOGE, Musk also pushed directly into the machinery of the federal government. Along with Peter Thiel and Palantir, Musk organized a “hackathon” to integrate citizen data across the IRS, Social Security Administration, and other agencies.

Using technology to reduce bureaucracy can sound perfectly reasonable. Faster services, less paperwork, fewer agencies clanking along on ancient software — who wouldn’t want that?

The concern is where all that power goes once the machinery is connected.

A more centralized government is more efficient. It also creates larger levers for fewer hands to pull, including hands the public never voted for.

—

Peter Thiel

After co-founding PayPal with Musk, Thiel became Facebook’s first outside investor in 2004, then co-founded Palantir, a government software contractor focused on defense, intelligence, and surveillance technology. He also runs investment firms that back financial technology and weapons systems, including battle-drone companies like Anduril.

So, again, a normal little career path.

Thiel is also deeply interested in biblical ideas like the Antichrist and the “katechon,” the holy power which can restrain the Antichrist. He has described the Antichrist as a “luddite who wants to stop all science,” naming Greta Thunberg as a possible henchman of the antichrist. Our era needs a katechon to keep the Antichrist at bay, he says.

That concern may sound, at first, like concern for human civilization. But in a recent interview, Thiel hesitated when asked, “You would prefer the human race to endure, right?” He then described his own transhumanist leanings.

Thiel also played a major role in J.D. Vance’s rise. Vance first met Thiel when Thiel gave a lecture at Yale Law School. Thiel later hired Vance at Mithril Capital, then backed Vance’s Cincinnati venture firm Narya, and helped bring him into the Silicon Valley political network.

Thiel became Vance’s most important early political patron, donating $15 million to support his 2022 Senate run and bringing him to his first Mar-a-Lago meeting with Trump in 2021.

As vice president, Vance has used his office to criticize Europe’s AI regulation movement and champion a more tech-centered, anti-regulatory vision for the federal government.

Musk has also championed Vance as the “Best VP ever and our future President.”

So the question is fairly simple: are Musk and Thiel only trying to make government work better — or are they helping prepare a much bigger disruption of government itself?

Thiel hasn’t been shy about an answer.

Got it, loud and clear.

—

Alex Karp

Alex Karp is the CEO and co-founder (along with Thiel) of Palantir, a roughly $400 billion company built largely around AI-powered “war-fighting” software for the U.S. government, NATO, and allied governments. The company has also been contracted by ICE to build software that helps track citizens, and Karp has said he wants Palantir to play a larger role in domestic crime prevention.

So the little software company has grown into something more like a nervous system for the national security state.

Karp’s new book, The Technological Republic, was recently condensed into a 22-point manifesto published on Palantir’s X account, which you can read here.

Karp has said there’s a two-thirds chance Palantir is involved in any major government foreign-warfare headline.

Trump recently praised Palantir for its role in the Iran attacks, posting: "Palantir Technologies (PLTR) has proven to have great war fighting capabilities and equipment. Just ask our enemies!!!."

The company’s stock value began rising immediately after Trump’s post.

—

Larry Ellison

Larry Ellison is one of the Top 10 richest men in the world. His wealth comes from Oracle, a major federal data contractor.

Last year, Trump’s “OneGov initiative” and “American AI initiative” deepened the government’s partnership with Oracle, with the company offering 75% discounts to federal agencies for AI integration, database management, cloud storage, and analytics.

Earlier this year, the Trump administration forced TikTok to form a new company for its U.S. users, called TikTok USDS. Oracle received a 15% ownership stake.

After the takeover, TikTok’s privacy policy was changed to “permit the company to track users' precise location, monitor their interactions with artificial intelligence, and provide collected information outside of TikTok.”

—

Here Comes the Technocracy

So what happens when we put all these pieces together?

The tech elites already control much of the public digital ecosystem. They make our devices, store our data, host our websites and emails, shape our media feeds, and track huge amounts of online activity.

They are also deeply embedded in government, public infrastructure, military contracting, law enforcement, border security, and national cybersecurity.

Thiel’s investment world has helped bring military and AI weapons companies into the mainstream, including Stark, Quantum Systems, and Anduril. “The natural fix is to put AI on the drones and turn these into more autonomous weapon systems,” he said in an interview.

Meanwhile, Musk is preparing to manufacture humanoid robots at scale.

We’ve all seen sci-fi movies, right?

Military systems, financial systems, communication systems, identification systems. These are some of the major levers needed to run a modern country. And the same small world of tech companies, defense contractors, investors, and political allies is moving closer to all of them.

Under Trump, the door opened wider for efforts to connect federal agency data and software infrastructure into a “Mega API” — a more centralized system for citizen data and government processing.

The next pressure point is data center security. If foreign or domestic actors start attacking data centers, there will likely be a major push for more surveillance, more automation, and more militarized technology inside U.S. borders.

Add public anger, institutional distrust, political violence, and we may find people choosing sides in a civil conflict — Musk’s "existential war” — that leads to government disruption.

When U.S. internal conflicts start to psychologically overwhelm the public, we can anticipate Musk & Co. launching their “network states”.

You know what those are, right?

You can read about them at thenetworkstate.com, but the short version is this: network states are proposed non-geographic political entities, formed by people around the world and organized through digital infrastructure.

This is what Thiel is pointing toward when he talks about building “new countries, new governments.”

And how would global citizens of a Musk-shaped network state communicate, verify identity, move money, pay fees, vote, or conduct daily life?

Through an “everything” app like Musk’s X.

Banking is one of the easiest ways to normalize I.D. verification. Once identity, payments, communication, and social life are tied together, digital voting and governance start to look much more technically possible.

Musk isn’t the only one thinking along these lines. OpenAI founder Sam Altman’s “World” project uses eye-scanning “Orb” technology to verify a person’s “Global I.D.” The project is being beta-launched in poorer countries by offering small cash bonuses for enrollment.

There’s much more to say here, but the main point is simple:

Of course civilization in the year 3000 A.D. will look very different from civilization today. Technology, science, war, politics, natural disasters, economics, and spiritual beliefs will keep changing the world.

But we’re already halfway there.

The richest men in the world — the ones shaping the internet, robotics, satellites, finance, AI, defense technology, and national politics — are moving quickly and they are succeeding. Public resistance has been pretty thin.

So as we think about the future and the policies we want to support, one big question should stay on the table:

How much power do we want centralized in the hands of tech elites?

This will be the defining question of the next 10 or 20 years.

There are plenty more dots to connect: the SEC’s recent push toward blockchain-based stock ledgers, Thiel’s new crypto bank, Iran’s demand for crypto payments in the Strait of Hormuz, the mental health crisis, transhumanism, Palantir/Microsoft/Google involvement in U.S. agriculture — the whole glowing infrastructure of the digital future.

And then there’s the Giant Mecha-Elephant in the room: artificial intelligence.

The dangers and promises of AI deserve their own documentary. In fact, a pretty good one was released last month:

There’s a reason so many science-fiction stories eventually wander into AI, and why so many people imagine it leading either to utopia or dystopia. There’s also a reason AI is driving a major share of new data-processing demand.

It’s a big deal.

Now, let’s turn to the mecha-elephant in our backyards: Data centers.

The Data & AI News Cycle

You’ve heard of the Dunning-Kruger effect?

This is the typical evolution of how we form opinions about an issue.

Early on, we learn a few things and start connecting the dots in familiar ways. This is good, that is bad, these people are right, those people have lost the plot.

Then a contrary detail appears. Someone else connects the dots differently. Often, our first instinct is to brush it aside: “Psshhht — nonsense. Fake news. Misinformation.”

News channels — corporate and independent alike — understand this human weakness. Ambiguity makes people anxious. We want our news reporting to be confident and simple.

The easiest story to sell is the old frontier tale: us good guys on one side, villains on the other. Media producers and politicians have been working that “Gunsmoke” model for many decades. Over time, it trains a culture to see every issue as a line in the sand — another battle in an existential culture war.

But data centers don’t fit as neatly into the traditional narrative. Rightwing outlets are disagreeing with each other.

The cover story on this month’s edition of New American is “The AI Prison Planet: The Dangers of AI and Data Centers, and Why Americans Are Turning Against Them”. Recent headlines from Blaze Media include “It’s past time for the government to rein in AI” and “The 3 biggest lies justifying massive AI data centers”

But over on FOX News and the Daily Wire, the headlines read “Agitators united by Chinese money, hate for America, target data centers, experts warn” and “Left-Wing Foreign Billionaires Fund Groups Trying To Cripple AI Infrastructure.”

Arizona’s conservative Goldwater Institute has been warning that our state’s “data center future” is being threatened by “overly restrictive zoning, unpredictable permitting, and policy driven by misconceptions.”

Meanwhile, most left-leaning outlets have leaned into anti-data center narratives. The headlines are nonstop. My friends at Tucson Agenda just published an anti-anti-anti-data center editorial:

This pattern is reversed in the government. Democratic officials and candidates have begun campaigning on mistrust of AI and data centers as a powerful swing voter issue, but at the top of the party, the winds shift direction: officials see data center property taxes as a way to fund infrastructure, and AI as a tool for improving government operations.

• “All in on AI: Shapiro, Amazon announce $20B AI investment in Pennsylvania.”

• “Governor Newsom partners with world’s leading tech companies to prepare Californians for AI future.”

• “Gov. Whitmer on Stargate Announcement from OpenAI, Oracle, & Related Digital Creating Thousands of Jobs, Largest Investment in Michigan History.”

• “New Jersey Gov. Phil Murphy Teams With NVIDIA to Advance AI.”

• “Governor Hochul Delivers Artificial Intelligence Training Tool to the New York State Workforce.”

On the Republican side of the tracks — with Florida as a major exception — top officials across the country and Arizona are more unified in their enthusiasm about data centers and AI development.

• Arizona legislators launch “initiative to strengthen the state’s leadership in AI and the digital economy.”

• Arizona Senator T.J. Shope wants data centers in rural Arizona, resists Gov. Hobbs push to end tax incentives for Big Data.

• Arizona Rep. Carbone says “Stop demonizing data centers — Arizona needs them to grow.”

• Governor Abbott and Google Announce $40 Billion Investment In Texas, Abbott says we “must ensure that America remains at the forefront of the AI revolution.”

• Mississippi Gov. Tate Reeves celebrates xAI’s “record-shattering $20 billion investment” in new Mississippi data centers.

Thankfully, Ground Party Papers subscribers seem tolerant of my devotion to honest and deep reporting, even if it doesn’t deliver comforting Sith-vs-Jedi narratives.

As we’ll see next, an honest reckoning with data centers is complex, high-stakes, and can cause indigestion.

Inside the Machines

Tech elites built the machines, shaped the algorithms, and made fortunes from the system. But their wealth did not appear out of some dark rift in the universe. It came from us — from the daily demand for data that now runs through our jobs, schools, businesses, entertainment, public infrastructure, and private lives.

Social media, streaming video, cloud storage, email, web services, online shopping, database hosting, internet searching and browsing — this is the bulk of ordinary data demand.

Here in America, we sit at the top of the techno-global food chain. Most of us enjoy the benefits without thinking much about the hidden costs.

Then the costs arrive in our backyard.

A mega dairy. A lithium mine. A smelter. A data center.

I’ve joined, and sometimes helped start, fights against big developments like these. But I try to do serious research first. Here’s what I’ve learned.

Data center demand has been climbing since the 90s, and it accelerated when everything started moving onto the “Cloud.”

Which human activities demand the most data? That question is harder to answer than it should be, because data center owners do not exactly invite the public under the hood. But researchers have found other ways to track the traffic.

Looking only at data moving between data centers and user devices, more than half of data demand is video streaming, mostly on YouTube, Netflix and social media. YouTube alone receives 500 hours of video uploads every minute, and users stream roughly 700,000 hours of video every minute.

A person can now watch a 4K movie on Disney+ while scrolling Facebook reels on the couch. Our cars, cameras, appliances, and spare moments are all being pulled into the Cloud.

Prediction: “Data conservation” is going to become part of the public discourse, just like “water conservation”. Expect articles with titles like “5 ways to reduce your daily data consumption”.

Server-to-device traffic is only part of the picture. Most data center activity happens server-to-server, inside the machinery of the internet.

When you search something on Google, your device sends a tiny amount of data. The results you receive, mostly text, are also small. The heavier action happens in between, as Google’s servers talk with other servers, indexes, databases, and ranking systems to decide what to show you.

The same kind of hidden choreography is happening all day long with our internet-connected apps, browsers, and devices, at home and at work.

This makes it hard to draw clean lines between the functions data centers serve. It’s more like a living ecosystem of information and computation. Different reports slice it up in different ways.

Drawing from all the best sources I found, I made this graph of total data center demand — our best ballpark estimates.

Even if this graph is accurate, energy use adds another layer of analysis.

AI servers tend to do more internal processing and use more electricity than regular servers, so AI’s share of data center energy demand is greater than its share of data traffic.

Then the picture widens again. Once you account for video production and the screens we watch on — TVs, laptops, phones — streaming video uses more total energy than AI.

Another thing that gets proliferated in the information ecosystem is a link between AI and hyperscale data centers, like the Project Blue data center outside of Tucson.

Hyperscale data centers are like the industrial agriculture of data — they’re much bigger, owned by magnates, and are more economically efficient. The hyperscale boom began in the 2000s, and a 2016 report from Lawrence Berkeley National Lab correctly predicted that they’d keep expanding into the 2020s.

These hyperscale centers are increasingly “AI-capable” but they’re mostly built to power the cloud hosting and back-end processes of giants like Amazon, Google, and Microsoft. AI remains only 25% of new data demand.

We could delete AI from the planet and data centers would keep expanding, if a bit slower, because the rest of our digital lives keep asking for more data. Lawrence Berkeley National Lab published a second study in 2024 which illustrates data center demand growth from AI and non-AI services:

Verdict: The details are complicated, and the responsibility for data demand is widely shared.

Historically, data center construction clustered in a few hot spots, with Northern Virginia’s “data center alley” at the top. The region has more than 600 data centers, processes the majority of global internet traffic, and uses 25% of Virginia’s electricity.

As those hubs fill up, development spreads outward. There is also a national security concern in concentrating so much digital infrastructure in a few places.

Data demand is expected to triple over the next five years, and its footprint is expanding into backyards across the county.

So, chasing data centers out of our backyard just means they land in someone else’s.

Or in Earth’s backyard.

Jeff Bezos and Elon Musk both want the majority of future data centers to float in Earth’s orbit. The pitch is simple: in space, data centers could run on solar power, avoid land-use problems, and escape the protests of angry citizens.

Musk lays out the idea on the website of his newest company, Terafab.

In January, Musk asked the federal government for permission to launch one million data center satellites into orbit. His longer vision is even stranger: manufacture them on the far side of the moon, then fling them into orbit with a giant slingshot — much cheaper than rocket launches from Earth. You can watch the video at terafab.ai

It’s all a very strange and disorienting new world.

And it’s understandable that people fight local data centers. Our survival instinct is to not race into uncertainty at full speed.

Recently, Tucson tried to annex Pima County land so the large “Project Blue” data center would sit inside city limits and pay major property taxes. A grassroots campaign stopped the annexation, and local politicians changed their tune.

But the petition did not stop Project Blue itself. Construction continues on county land, and instead of using Tucson’s reclaimed wastewater as first planned, the project will use local groundwater.

Protests may annoy developers and slow things down, but they do not end the sprawl. These companies will find struggling rural towns that badly need property tax revenue and will say yes. It’s happening across the U.S. and around the world, including in Central and South America.

It will be harder to stop as President Trump, and perhaps a future President Vance, push AI data centers through “Project Stargate,” a new federal partnership with Oracle and OpenAI.

The project includes $500 billion for AI data center infrastructure, while the White House’s “AI Action Plan” aims to speed permitting by cutting “bureaucratic red tape.”

So we should ask a hard question: If we scare data centers out of our backyard, will we sleep better knowing they’ve just moved somewhere else?

“Out of sight, out of mind” may be the most dangerous attitude available.

Data centers built elsewhere — especially in space — move beyond local reach. What will you do when that technology is used for surveillance and domination? A garage full of rifles won’t mean much against slaughter drones or humanoid sniper robots controlled by AI data centers floating in space.

Maybe we’ve forgotten the old adage: Keep your friends close and your enemies closer.

But no one forgets this adage: If you can’t beat ‘em, join ‘em.

If resistance to the Network State means getting targeted by the richest people and most advanced weapons on Earth, plenty of people will join Team Technocracy and justify it as self-preservation.

While you chew on that consideration, the next factor to unpack is water.

Water, Energy & Money

Let’s zoom in on Cochise County.

Nearly all of our groundwater basins are strained to some degree. The Douglas and Willcox AMAs prohibit new agricultural irrigation, but not new industrial water use like feed lots, mines, or data centers.

Presently, over 90% of groundwater use in the Douglas Basin is agricultural. In the Willcox Basin, it’s ~97%. So any real balancing plan has to include major reductions in agricultural water use.

Delay makes the damage worse. Residents lose wells. Earth fissures spread. Land subsidence permanently reduces aquifer storage.

It also hurts agriculture. As water drops, well pumps work harder, electricity costs rise, and narrow farm margins get thinner.

When people warned that groundwater regulation would hurt agriculture, local farmer Ed Curry answered: “If we keep doing what we’re doing, we’re going to face the same consequences.”

He’s right – with or without regulations, growers will get less water. The real choice is between reducing agricultural pumping while many wells still reach the water table, or waiting until only corporations with 2,000-foot wells control the water.

Well – what can we do to balance our water budget while also bolstering local food security and local small family farms? That’s a hard question for another edition.

Today’s question is how much data centers could worsen the water problem.

Data centers generate heat, and cooling them consumes water. The three main cooling approaches are evaporative, non-evaporative, and direct-to-chip.

Evaporative cooling is the big on-site water guzzler. It relies on cooling towers where water evaporates and leaves mineral buildup that has to be flushed, producing wastewater called blow-down. Blow-down can contain anti-corrosive chemicals, some poorly regulated, which could cause environmental problems when handled irresponsibly.

So evaporative cooling introduces several big concerns at once.

Non-evaporative cooling, often called “dry” or “closed-loop” cooling, is the popular low-water-use option on site. It uses more electricity, and power generation also consumes water, but its total water footprint is still much smaller than evaporative.

Direct-to-chip cooling (confusingly, also called “closed-loop” sometimes) has even less water and energy use, but it's new tech and less adopted.

By my research, here is the rough annual water footprint per 500,000 square feet of air-conditioned space, including water consumed for electricity generation by the three cooling systems:

• Evaporative: 1,804 acrefeet

• Non-evaporative: 200 acrefeet

• Direct-to-chip: 77 acrefeet

Can data centers significantly affect rural groundwater? With evaporative cooling, yes. In the Douglas Basin, another 1,804 acre-feet of demand would increase total pumping by about 2.9%.

A non-evaporative system would add about 0.3%. Direct-to-chip would add about 0.1%.

And with non-evaporative cooling, that water usually comes through electricity generation, not the local aquifer, unless the data center produces power on site.

If it uses on-site solar or gas, the water footprint drops close to zero.

I don’t think we can afford data centers with evaporative cooling. But given Arizona’s water problems, should we allow any data centers at all?

The answer depends on how you feel about… 

Money

If you replaced all of Cochise County’s farmland with data centers, Cochise County would become a dystopian hellhole — but an insanely rich hell hole with a budget in the billions instead of millions, and much less groundwater overdraft.

I, for one, don’t want to live in a rich hell hole, even if it eliminates residential property taxes and covers the cost of private wells for everyone who wants one.

Like most of you, I love rural Cochise County’s scenic views and quiet nights more than riches. I’d be — annoyed, to put it lightly — if even one noisy data center landed in my backyard.

Elected officials in rural areas understand this mindset. Some share it. So why aren’t they protesting data centers with the same fervor as many of their constituents?

Let’s do some role-playing.

Imagine you’ve just been elected as a Cochise County Supervisor.

The county has a slow economy and a stagnant population base. Federal grant funding for rural communities is starting to disappear.

Meanwhile, voters want jobs, better roads, emergency response, flood control, and basic public services across a large, spread-out, mostly unincorporated county.

County government also has to pay employees, update systems, maintain buildings, process lawsuits, and coordinate with other agencies at the local, state, and national level.

That all costs money. Inflation pushes prices up. And citizens do not want tax hikes.

Dang.

You’re not a techno-fetishist and you’re not looking for backroom money grabs. But a data center developer shows up with a project carrying a massive property valuation. With that tax revenue, the county can finally fill the budget holes from the demands of residents and department heads who consistently need project funding.

If you’ve ever looked at the finance books, you know Cochise County already runs a tight budget. So, what to do?

It’s not a very fun role-playing scenario.

Of course, there are plenty of reasons to shake our fingers at elected officials (like some of our county supervisors’ interest in bypassing or deleting Open Meeting Laws).

But in this battle, the supervisors are not a useful target any more than you or I are. They’re just a minor cog in a much bigger machine, trying to solve wicked problems that no one really knows how to solve.

I sure as heck wouldn’t want their job.

Speaking of tricky policy problems, we gotta talk about…

Energy

Like data demand, humanity’s energy demand keeps rising.

If supply falls behind, energy gets more expensive. When energy is more expensive, literally everything else is more expensive.

Data centers, streaming video, and AI all add pressure. Increasingly, data centers are building on-site power, using solar, natural gas, or small nuclear reactors.

Anyone who reads local Facebook groups knows Cochise County has a small but loud anti-solar crowd. I get that. I don’t want a solar farm next to my home either.

But if we don’t keep up with energy demand, prices rise for everyone, including farmers who need electricity to pump water and grow food on narrow margins.

So we still have to choose some flavor of energy production. Every energy source has tradeoffs, but if we care about water conservation, air quality, and supply chain vulnerabilities, the best viable option in the desert is solar.

There’s more to say about energy — too much more — but I needed to mention solar to set up a punchline at the end of this story.

A Path to Power

Finally, I’ll offer my three cents on how Cochise County could shape data center policy. Then I’ll add some bonus ideas for readers interested in “collapse” scenarios — global conflict, culture-war breaking points, AI apocalypse, climate change, or whatever flavor of doom keeps you up at night.

Disclaimer: the global AI and data center story is changing daily. After three months of research, I have some ideas, but I reserve the right to change my mind as the world keeps turning.

Barring some unexpected event — maybe the arrival of extraterrestrials, nuclear armageddon, or an anti-tech messiah uniting humanity — data centers are here to stay, just like computers and the internet.

But that doesn’t mean we just roll over when they show up in our backyards.

Here’s how Erin Brockovich recently advised citizens facing local data center projects:

I agree, and it’s why I wouldn’t tell people to simply stop resisting data center development. If you don’t resist, the developers won’t be forced to reduce their resource consumption.

Paul Hirt, a Portal resident and retired professor who specializes in natural resource management and energy policy, recently talked with the Arizona Republic about the first wave of resistance to data centers:

Along with public benefits, Cochise County needs strong guardrails on resource use. I’d suggest these conditions for data centers (and their onsite energy projects):

1. Data centers should be considerably set back from residences, unless nearby residents give signed approval.

2. Strong limits on noise pollution, including tonal sounds and decibel caps beyond a set distance.

3. Data centers cannot use evaporative cooling.

4. Data centers in AMAs must buy or lease grandfathered water duties equal to 200% of their total water footprint, including off-site energy production, and retire half of the duties.

5. At least 75% of a data center’s energy demand must come from private power generation or subsidized new grid energy.

6. Data center owners must have bond contracts covering full decommissioning costs.

7. Projects must complete ADEQ and EPA environmental impact surveys and release the results before county approval.

8. A share of tax revenue must go into a county fund for residents affected by unexpected problems caused by data centers.

Regarding the county’s initial memorandum on data center approval, I share many criticisms already made by Cochise residents.

Here are my line-by-line responses.

• “Data Centers shall be allowed by Special Use Authorization only in the HI, LI, RU, and GB Zoning Districts.”

Given the risk of noise pollution, data centers should be limited to Heavy Industrial (HI) zones.

• “Additional Special Use Application Requirements. … (d) Noise impact analysis, if requested by the County, based on site location and potential noise-generating equipment.”

Noise analysis must be mandatory — not stipulated — for every proposal.

• “Potable water shall not be used for routine cooling or humidification unless the Applicant demonstrates, through the Water-Use Plan, that non-potable or reclaimed water is not reasonably available or feasible.”

For rural sites, reclaimed water usually won’t be available, leaving only groundwater sources. In AMAs, any new groundwater demand should be offset by buying and retiring water duties from corporate agriculture in the same basin.

• “Applicant shall provide … projected electrical demand…”

Data centers should build private power instead of adding major demand to the grid.

These are strict requirements. Data centers deserve strict requirements. They can afford them.

I want to especially insist on noise regulations for data centers. Past resistance has pushed developers to innovate technologies that decrease their energy and water demands — they need to be similarly pushed to develop quieter technology as well. Might as well be Cochise County that does the pushing — we’re good at it.

I’ll start wrapping this up, but I promised a few less-explored perspectives on data centers.

The first is solar energy.

If we think eight steps ahead, decades into the future, we may be glad to have plenty of solar panels nearby.

In any serious collapse scenario, most of the hundreds of thousands of people in the Arizona-Mexico borderlands would have no power, no fuel, no refrigeration, and no way to pump water. The Apache Station would have no coal to burn. The masses will need to get power somehow.

An off-grid home with panels and batteries might look self-sufficient at first, but it would quickly become a target.

The threat would mostly come from the largest organized groups, not lone bandits. Law enforcement, military units, cartels, militias, or other armed factions would be hunting for functional infrastructure. If they wanted your panels, batteries, fuel, or guns — or your entire home — they would have the numbers to take them.

Commercial-scale solar provides an out. More panels spread across the region could power local microgrids and remove the target from off-grid homesteads.

If data centers are required to build their own solar power, that means more collapse-resistant energy infrastructure in the county.

Data centers also bring another strange possibility. Each one is a hardened structure with power systems, UPS batteries, backup generators, cooling equipment, and security infrastructure. In a collapse scenario, that concentrated capacity could become a community resilience node: communications hub, emergency clinic, cold-storage depot, charging station, dispatch center, or food distribution point.

But I’m not here to preach collapse and doom, nor techno-optimism — just food for thought.

I don’t know what will happen. I’m trying to research the issues, think through as many angles as I can, and see what questions remain.

If tight restrictions mean no data centers get built in Cochise County, I won’t shed tears. If they do get built, I’ll look for the silver linings.

Whatever happens with data centers, our best work is still close to home: knowing our neighbors, buying local when we can, growing food when we can, showing up before the big decisions are already made, and gathering together face to face.

Host the potlucks, organize the shindigs. Get the ranch folks, church folks, hippie folks, veteran folks, teacher folks, business folks, and weird desert hermits in the same room once in a while.

Cochise County’s deepest resilience will come from its people.

Community is our most important natural resource.

Have something to say about this wild and complicated story?

The comments section is open!

If you enjoyed this newsletter, forward it to a friend.

And…

Thanks for reading!