April 28, 2026
The AI economy needs physical infrastructure. The question is who hosts the costs — and who captures the upside.
AICV Civic Intelligence Report No. 001 AI Infrastructure · Data Centers · Utilities · Land Use · Community Impact Published: April 2026 Region: Coachella Valley / Imperial Valley / Southern California
This report examines the proposed data center activity in Coachella through a broader civic intelligence lens: AI infrastructure, land use, utilities, water, ratepayer risk, community benefit, and regional economic development.
It is not an anti-technology argument.
It is not a pro-developer argument.
It is a public-interest question:
If the AI economy needs physical infrastructure, why are certain communities asked to host the costs while distant companies capture most of the upside?
Before we get into the good, the bad, and the ugly of data centers, I want to start somewhere simple.
If you live in Coachella, or anywhere in this region, and you are pushing back against a proposed data center in your community — I’m with you.
You have every right to ask questions. You have every right to protest. You have every right to demand answers before hundreds of acres of land, grid capacity, water, and public trust get committed to a project most residents had no meaningful role in shaping.
That does not make you anti-technology. That does not make you anti-business. That does not make you anti-progress.
It means you understand something a lot of people in the AI economy would rather not talk about: the cloud is not actually in the cloud.
It lives somewhere. It uses land somewhere. It pulls power from someone’s grid. It needs water, cooling, roads, permits, emergency services, and political permission.
And increasingly, the places being asked to host that infrastructure are not the places capturing most of the wealth it creates.
That is the real question here.
If the AI economy needs physical infrastructure, why are certain communities asked to host the costs while distant companies capture most of the upside?
That is the frame for this report.
I’ve been writing about AI, data centers, and the physical infrastructure behind the digital economy for the last couple of years. And personally, no, I’m not excited about the idea of turning parts of the Coachella Valley into server-farm territory. I don’t think every desert community with land, sun, and grid access should automatically become a sacrifice zone for the next phase of the internet.
But I also don’t think this conversation is as simple as “data centers bad.”
Data centers are not new. They have powered the internet for decades. Every app, every search, every streamed song, every AI chatbot, every cloud backup, every DoorDash order, every Instagram reel — all of it runs through buildings full of servers somewhere.
The difference now is scale.
AI needs far more computing power than the old internet did. That means more data centers, more electricity, more cooling, more land, and more pressure on communities that may not have asked to become the physical backbone of the AI boom.
So before anyone gets sold a glossy rendering, a jobs promise, or a vague line about “economic development,” we should slow down and ask better questions.
Why data centers? Why now? Why here? Who benefits? Who pays? Who decides? And what would a fair deal actually look like for the people who live here?
That is what this report is about. Not a rant. Not a press release. Not a blanket rejection. Just the good, the bad, and the ugly of what happens when the AI economy comes looking for land in the desert.
According to local reporting, the proposed Coachella project is a technology campus near Avenue 52 and Fillmore Street. The initial proposal involves roughly 240 acres, but city records reported by KESQ indicate the campus could eventually grow to as much as 450 acres with six data centers.
That is not a small project. That is not a few server racks tucked into an industrial park. That is a major land-use decision with long-term implications for energy, water, traffic, noise, agriculture, environmental review, and the future identity of the east side of the region.
Residents have already shown up in significant numbers. More than 100 people protested at a recent Coachella City Council meeting, with concerns focused on environmental impact, water, electricity, and transparency. KPBS also reported that residents are calling for a moratorium, while city officials have said the project has not yet been approved and would require an environmental impact report.
That distinction matters. As of publication, the project is not approved. But the conversation is very much underway.
And once a project like this gains momentum, the public often ends up debating details after the big direction has already been quietly shaped. That is why the process matters before momentum hardens into inevitability.
One of the biggest tricks of the digital age was branding physical infrastructure as something invisible.
“The cloud” sounds soft. Light. Effortless. Borderless.
But the cloud is not a cloud. It is a building. Often many buildings. Usually windowless. Usually loud. Usually power-hungry. Usually surrounded by fencing, generators, substations, cooling equipment, backup systems, and security.
The internet was never weightless. We just outsourced the weight.
First it was search, email, streaming, e-commerce, social media, mobile apps, and cloud storage. Then came crypto. Now AI.
Every time you ask ChatGPT a question, generate an image, summarize a document, run a model, or automate a workflow, somewhere a machine does work. That work requires compute. Compute requires electricity. Electricity requires infrastructure. Infrastructure requires land. And land belongs to real places.
So when we talk about AI, we are not just talking about software anymore. We are talking about a new industrial footprint. That is the part the technology industry often prefers to skip.
AI may feel magical on your phone. But on the ground, it looks like substations, transmission lines, backup generators, water demand, cooling systems, and zoning hearings.
Welcome to the physical internet.
The timing is not random.
AI models need massive amounts of computing power to train and run. The bigger the models get, the more infrastructure they require. The more people use them, the more demand grows. The more companies race to build AI into everything, the more pressure there is to expand the data center footprint.
The U.S. Department of Energy says data center deployment, partly driven by AI applications, is now a significant factor in near-term electricity demand growth. EPRI estimates that data centers could consume up to 9% of annual U.S. electricity generation by 2030, up from 4% of total load in 2023.
Pew Research Center found that the next wave of data centers is also shifting geographically. While most existing U.S. data centers are in urban areas, 67% of planned data centers are in rural areas, and 39% are planned for counties that do not currently have one.
That is the new map.
AI companies need land, power, speed, and political permission. Rural and semi-rural communities increasingly look attractive because they may have larger parcels, fewer immediate neighbors, cheaper land, and less organized resistance — at least at first. That is why communities like ours need to understand the playbook before the playbook gets used on us.
The Coachella Valley checks a lot of boxes.
We have land. We have proximity to Southern California. We have existing energy infrastructure. We have intense sunlight and a renewable energy narrative. We have cities hungry for economic development. We have agricultural land under pressure. We have communities that have historically been asked to absorb regional burdens without always receiving regional benefits.
That last part is important. Because this is not just a technical question. It is a power question.
Where do these projects get placed? Which communities are expected to say yes? Which communities get promised jobs? Which communities get told the benefits are coming later? Which communities get the renderings, and which ones get the noise?
As one Coachella resident put it in local coverage, someone made a choice to place this proposal where it is — not in a wealthier part of the region.
That is not a side comment. That is the geography of power showing up in real time.
No.
The Coachella Valley does not need a local data center in order to use AI, build AI-enabled businesses, train workers, modernize local institutions, or participate in the next phase of the digital economy.
That distinction matters.
AI services already run across distributed cloud infrastructure. Cloud providers operate networks of regions, availability zones, data centers, fiber, and edge infrastructure around the world. Users and businesses do not need a data center in their own city to access AI tools, build AI workflows, or create AI-native companies.
So the argument for local data centers should not be framed as: accept this infrastructure or be left behind. That is a false choice.
The real question is different: if a data center is proposed here, does it create enough local value to justify the local burden? That is the bar.
Not whether the Coachella Valley believes in AI. Not whether the region wants to modernize. Not whether small businesses, students, cities, nonprofits, hospitals, hotels, farms, and entrepreneurs should learn how to use these tools. They should. They can. They already are.
But participating in the AI economy and hosting the physical infrastructure of the AI economy are not the same thing.
A region can become smarter without becoming a server farm. A workforce can become AI-ready without putting hundreds of acres into data center use. A startup ecosystem can emerge here without requiring hyperscale compute to sit next door.
So if a data center comes to the Coachella Valley, it should have to stand on its own civic and economic merits — not on the fear that saying no means missing the future.
A data center proposal is never just a land-use proposal.
It is a utility deal. It is a power deal. It is a water deal. It is a substation deal. It is a ratepayer deal. It is a governance deal.
And that is where this gets complicated in the Coachella Valley.
Because this region already lives inside a layered infrastructure story. We have solar. We have wind. We have agricultural water systems. We have golf courses. We have luxury developments with man-made lagoons. We have farmworker communities on the east side where safe drinking water has not always been guaranteed. We have multiple electric and water agencies serving different parts of the region. We have summer peak demand, extreme heat, and a grid that was not originally designed for the AI economy.
So the question is not simply whether a data center can be plugged in. The question is what has to move, expand, bend, or be paid for so it can be plugged in.
That may mean new substations. It may mean transmission upgrades. It may mean backup power. It may mean water infrastructure. It may mean new interconnection agreements. It may mean special tariffs. It may mean a renewable energy package that sounds clean on paper but still requires land, storage, permitting, and grid capacity.
This is why utilities are not side characters in the story. In the Coachella Valley, the regional picture includes electric providers, water agencies, wastewater systems, irrigation infrastructure, substations, transmission planning, conservation mandates, groundwater management, and long-term climate pressure.
So the public should not only ask whether the city supports a project. The public should ask what the regional utilities know, what they have modeled, what infrastructure would be required, and what protections exist for existing residents and businesses.
Water is even more sensitive.
The region already makes hard choices about water. CVWD says 54.5 golf courses out of 105 in its service area use a nonpotable water source, and that 3.4 billion gallons of wastewater are recycled yearly at two wastewater treatment plants. CVWD also says more than two-thirds of local farmland is irrigated partly with Colorado River water delivered from the Coachella Canal.
Then there is the symbolic layer. Cotino, the Storyliving by Disney community in Rancho Mirage, is a 618-acre luxury residential development with nearly 2,000 planned homes and a 24-acre artificial lagoon. The point is not that Cotino is the same as a data center. The point is that in the desert, every major project that markets abundance becomes part of a larger public story about who gets water, who gets infrastructure, and who gets asked to wait. That story lands differently on the east side.
In parts of the Eastern Coachella Valley, safe drinking water has not always been a given. In 2024, the EPA announced it had overseen the restoration of safe drinking water to more than 900 residents living in 20 mobile home parks within the boundaries of the Torres Martinez Desert Cahuilla Indians Reservation, where drinking water had previously contained unsafe arsenic levels.
So when residents hear that a massive new AI infrastructure project may need power, water, cooling, substations, and public approvals, they are not being irrational. They are doing the math from lived experience.
The real review has to include the full utility ledger: what power is needed, what water is needed, what substations are required, what upgrades are triggered, who pays upfront, who pays over time, who gets protected from cost shifts, who benefits if utilities expand, and who is left with the infrastructure if the project underperforms, changes owners, or becomes obsolete.
The AI economy may be digital. But its infrastructure bills are local.
Let’s be honest. There are potential benefits.
Data centers can generate property tax revenue. They can create construction jobs. They can bring investment into cities that need a stronger commercial tax base. They can help fund services if the deal is structured properly.
And yes, the AI economy does need infrastructure. We cannot say we want the benefits of modern technology while pretending the physical systems behind it do not exist.
There is also a version of this where a data center becomes more than an isolated server warehouse. A serious, well-negotiated project could include workforce training, local hiring commitments, energy investments, public infrastructure upgrades, community benefit agreements, environmental protections, and direct funding for the neighborhoods absorbing the impact.
The Brookings Institution has argued that AI-era competition for mega-sites, grid access, and permits gives regions new leverage — and that regions should ask for negotiated co-investments that can anchor more durable local prosperity.
That is the optimistic version. A data center could be part of a larger regional modernization strategy. But only if the region negotiates like it understands its own value.
Here is the more common version.
The construction jobs are real, but temporary. The permanent jobs are usually much smaller than people imagine. Data centers are not factories in the old-school sense. They are capital-intensive, not labor-intensive. That means most of the economic upside often flows to the owners, operators, vendors, financiers, and cloud customers — not necessarily to the host community.
Brookings put it plainly: the standard data center development model tends to involve speedy dealmaking and opaque negotiations that deliver short-term construction jobs and revenue, but relatively little durable local economic upside.
That is the part every city council needs to sit with. Because “economic development” is not a magic phrase. A project can be large without being transformative. A project can be expensive without being generous. A project can increase a city’s tax base while still leaving residents with higher bills, strained infrastructure, and no meaningful pathway into the economy being built around them.
That is the bad version: the community hosts the machine, but never gets invited into the future it powers.
The ugly is when the process gets rushed, obscured, or overwhelmed by outside interests.
In Imperial County, officials recently advanced a proposed 950,000-square-foot data center complex near the city of Imperial amid fierce opposition. KPBS reported that residents opposed to the project waited outside the county administration building, access to the meeting became a major flashpoint, and sheriff’s deputies removed at least one organizer from the room. KPBS also reported that the project still lacked water and energy contracts at the time it advanced.
That should make everyone pause. Because this is not just about whether a project can be drawn on a map. It is about whether the public process is strong enough to withstand pressure from developers, unions, consultants, landowners, utilities, and elected officials who may all have different incentives.
If residents feel locked out, ignored, or handled, trust collapses. And once trust collapses, every technical answer sounds like spin.
That is the ugly part. Not just the data center. The feeling that the decision has already been made.
There is another layer here that makes this even more complicated.
The companies driving AI growth know the economic model is changing. OpenAI’s own industrial policy paper warns that AI may expand corporate profits and capital gains while reducing reliance on labor income and payroll taxes. The paper argues that policymakers may need to rebalance the tax base toward capital-based revenues, including higher taxes on top capital gains, corporate income, or sustained AI-driven returns.
Let me translate that. Even the companies building the AI economy understand that wealth may concentrate around capital, not labor. That means fewer traditional jobs may support more corporate profit.
That also means communities hosting AI infrastructure should be very careful about accepting old economic-development language at face value.
Because if AI is going to make some companies historically wealthy while reducing labor’s share of the upside, then host communities need a different deal. Not charity. Not sponsorship. Not a ribbon cutting. A real deal.
If the AI economy depends on physical communities, then physical communities deserve enforceable upside. That could mean local energy protections, community benefit agreements, training pipelines, local ownership structures, guaranteed public infrastructure investment, ratepayer protections, a regional AI workforce fund, or direct payments into affected neighborhoods.
But it cannot mean: “Trust us, this will be good for you.” We are past that era.
If Coachella, or any city in this region, is going to seriously consider data centers, the bar should be high. Not performatively high. Actually high.
At minimum, residents should expect:
A full environmental impact report. Not a shortcut. Not a workaround. Not a narrow review that ignores cumulative impact.
Transparent water and electricity analysis. How much power? From where? At what cost? Who pays for upgrades? What happens during peak summer demand? What happens to ratepayers?
Full utility clarity before political momentum builds. The relevant electric and water providers should not be treated as footnotes. Before any project advances, the public deserves plain-English answers about electric service, water service, projected demand, substation and transmission needs, cooling requirements, backup power, infrastructure upgrades, cost recovery, summer peak-load risk, ratepayer exposure, and who pays if the system has to expand.
A legally binding community benefit agreement. Community benefit agreements should be transparent, publicly available, legally meaningful, and include quantifiable details on jobs, tax revenue, workforce training, health, community benefits, infrastructure commitments, and cost-sharing.
Local hiring that means something. Not vague language. Not “local when possible.” Real targets. Real reporting. Real consequences.
A workforce pipeline for operational jobs. If the permanent jobs require technical training, then the developer should fund that training locally — through College of the Desert, CSUSB Palm Desert, local high schools, workforce boards, unions, and community organizations.
Ratepayer protection. If infrastructure upgrades are needed, residents should not quietly subsidize a private AI facility through higher bills.
A moratorium discussion if the city needs time. A pause is not anti-business. A pause can be responsible governance.
Public meetings residents can actually attend. Not during impossible hours. Not with technical jargon. Not after the important conversations already happened.
This is what good process looks like. And if a developer cannot accept that, that tells you something.
We should not fall for the idea that asking questions means we are against the future. The future should be able to survive a public meeting.
We should not fall for the idea that all investment is good investment. Some investment extracts more than it contributes.
We should not fall for the idea that because AI is important, every AI-adjacent project deserves automatic approval. Importance is not immunity.
We should not fall for the idea that data centers are either salvation or disaster. They are infrastructure. Infrastructure has tradeoffs. Tradeoffs require governance.
And we should not fall for the oldest trick in the economic development book: big number, small details. The big number is the headline. The details are where the community either wins or gets played.
I do not think the Coachella Valley should simply reject the AI economy. That would be foolish.
AI is going to reshape tourism, hospitality, healthcare, agriculture, real estate, education, media, government, and small business. The region needs to understand it, adopt it where appropriate, regulate it where necessary, and build new companies around it where possible.
But hosting the physical infrastructure of AI is not the same as participating in the AI economy.
A data center does not automatically make us an AI hub. A server farm does not automatically create startups. A substation does not automatically train students. A tax bump does not automatically modernize a workforce.
If we want the Coachella Valley to be part of the AI future, the better question is not just: “Should we allow data centers?”
The better question is: What would it take for AI infrastructure to produce durable local intelligence, local talent, local ownership, and local opportunity?
That is a much harder question. And it is the only one worth asking.
Because if the only thing we get from the AI boom is land consumption and grid strain while the real value flows elsewhere, then we did not build the future. We leased it our backyard.
This report will continue tracking whether the City of Coachella treats public engagement as a real decision-making process or merely an informational pressure valve.
It will track whether city leaders discuss a moratorium or other pause mechanism. It will track whether the relevant utilities provide clear, public, plain-English answers. It will track whether the developer offers an actual community benefit agreement or just soft promises. It will track whether the environmental review is treated as a serious process or a box to check.
And it will track whether the rest of the region understands that this is not just Coachella’s issue.
If this project moves forward, every city will learn from it. Developers will learn what kind of process works here. Residents will learn whether their voices matter. Elected officials will learn how much scrutiny the public is willing to apply. And the AI industry will learn what kind of region the Coachella Valley intends to be.
That is why this matters. Not because one project determines everything. But because first fights set patterns.
I am not here to tell anyone what to think.
I am here to say this: do not let anyone make you feel small for asking large questions.
Hundreds of acres is a large question. Energy demand is a large question. Water is a large question. Agricultural land is a large question. Ratepayer risk is a large question. The future of the AI economy is a large question.
And the people who live here are allowed to ask all of them. Especially before the machines arrive.
So participate. Ask better questions. Demand real answers. And remember: the cloud may be virtual, but the consequences are local.
Sat Singh is the founder of SunshineFM and AI Coachella Valley. He has been writing about AI, infrastructure, and the future of the Coachella Valley since before data centers became the local headline.