Erin Brockovich has a new target, and the tech media is eating it up.
With the launch of her crowdsourced AI Data Center Reporting map, the famous paralegal turned consumer advocate has turned her sights on the massive, windowless structures powering our digital lives. The narrative is as predictable as it is flawed: greedy hyperscalers are sneaking into rural America, secretly stealing water, bankrupting the local power grid, and leaving citizens with nothing but noise pollution and high utility bills. If you found value in this article, you should check out: this related article.
I have spent fifteen years managing infrastructure deployments and negotiating power purchase agreements for the world's largest cloud operators. I have seen companies blow millions trying to appease local activist groups, only to realize the core premise of the outrage was entirely disconnected from how the physics of a modern power grid actually operates.
The current hysteria surrounding data center secrecy is a classic case of fighting the wrong battle. The lazy consensus insists that transparency is the antidote. If we just map every server farm, force operators to disclose every gallon of water, and shining a giant spotlight on the tech industry, we can fix the problem. For another perspective on this development, see the latest coverage from Mashable.
This is a delusion. Data center secrecy is not a conspiracy to poison your well water; it is a fundamental design requirement of national critical infrastructure. Furthermore, the local panic over energy and water scarcity ignores the basic economic realities of utility engineering. If you want to protect American communities and build a resilient infrastructure, you need to stop demanding useless transparency and start rewriting the regulatory framework governing utility monetization.
The Myth of the Secret Server Farm
Let’s dismantle the premise of the Brockovich map immediately.
The idea that hyperscalers are building multi-gigawatt facilities in total secrecy is laughably absurd. You cannot secretly construct a three-million-square-foot industrial complex. You cannot secretly order 400-megawatt transformers that have a four-year manufacturing lead time. Every single one of these projects goes through months, sometimes years, of public zoning boards, environmental impact reviews, and county commissioner votes.
When activists complain about "secrecy," what they are actually angry about is non-disclosure agreements (NDAs).
During the site selection phase, a developer might operate under a shell company like "Project Thunder" or "Pluto LLC." Critics call this deceptive. In the real world, it is standard corporate procurement. If Microsoft or Google announces they are looking at 500 acres of farmland in rural Ohio under their real names, land speculation skyrockets overnight. The price per acre jumps from $5,000 to $100,000, rendering the project economically unviable before the first shovel hits the dirt.
More importantly, physical security dictates operational obscurity. Under federal guidelines, data centers house critical national infrastructure. They hold proprietary intellectual property, financial records, government cloud workloads, and health data.
I have designed facilities where the perimeter security rivals military installations. We do not put up neon signs saying "Google AI Cluster Inside" for the same reason banks do not display their vault combinations on the front window. Demanding that operators publish granular, real-time internal schematics or precise equipment layouts under the guise of "community awareness" is an open invitation to state-sponsored physical sabotage.
The Energy Puzzler Nobody Understands
The headline-grabbing statistic everywhere right now is terrifying: data centers could consume 9% of total US electricity by 2030. The immediate reaction from local communities is panic that their lights will go out or their household bills will double.
This completely misunderstands how grid infrastructure is financed.
Consider how a standard utility operates. A regulated utility makes money by spending capital on assets like transmission lines and power plants, and then charging a rate of return on those assets, approved by a public utility commission. For decades, electricity demand in the United States was entirely flat. Utilities had no incentive to build new, cleaner generation because they did not have the customer demand to justify the capital expenditure.
Enter the hyperscalers. Data centers represent a massive, predictable, 24/7 baseload. They are the ultimate anchor tenants for the modern grid.
When a technology company signs a power purchase agreement (PPA) for 500 megawatts, they are not stealing power from residential neighborhoods. They are providing the bankable financial guarantee that allows utility companies to build brand-new generation resources—predominantly utility-scale solar, wind, and advanced nuclear.
Look at the Ratepayer Protection Pledge signed in early 2026. The major players explicitly committed to funding new generation resources rather than cannibalizing existing local capacity. The narrative that data centers simply leech off the existing grid is an outdated myth. They are funding the transition to a modernized grid that would otherwise take decades to finance.
The Brutal Reality of Ratepayer Bills
But what about the recent stories of ratepayers getting stuck with infrastructure bills? In Maryland, residents faced a $2 billion grid upgrade bill linked to regional power transfers.
This is where the contrarian view must acknowledge the downside: the current system is broken, but not for the reasons Brockovich claims.
The problem is not the presence of the data center. The problem is antiquated utility regulatory structures that allow power companies to socialize the cost of regional transmission upgrades across the entire ratepayer base instead of directly allocating 100% of those specific costs to the industrial user.
If a data center requires a new high-voltage transmission line to pull power from three states away, the utility company should legally be required to charge that specific operator an un-subsidized tariff that covers the entire asset lifecycle. Instead, outdated regional transmission organization (RTO) rules often dictate that grid reliability upgrades are shared among all consumers.
Blaming the tech company for utilizing a legal utility tariff is pointless. It is a failure of local Public Utility Commissions (PUCs) to update their regulatory frameworks for the modern industrial era. If you want to protect citizens from rate hikes, stop protesting outside a construction site. Start packing the public hearings at your state’s utility regulatory commission and demand a complete overhaul of industrial tariff structures.
The Water Scare vs. Engineering Physics
Next to power, water is the primary battleground. The Brockovich platform highlights community complaints about data centers sucking local aquifers dry. The standard talking point is that a large facility consumes millions of gallons of water per day, comparable to a small city.
This argument is stuck in 2012.
Early data centers relied heavily on evaporative cooling systems because they were cheap and highly efficient at lower ambient temperatures. But the industry has shifted rapidly due to sheer operational risk. Relying on millions of gallons of municipal water is a massive liability for a multi-billion-dollar facility. If a local drought hits and the municipality cuts off industrial water access, the data center goes dark.
Because of this risk, modern AI clusters are moving aggressively toward closed-loop liquid cooling systems or direct-to-chip cooling architectures.
$$PUE = \frac{\text{Total Facility Energy}}{\text{IT Equipment Energy}}$$
To optimize Power Usage Effectiveness ($PUE$), engineers now use closed-loop configurations where the same water is recycled indefinitely through a sealed system, using air-cooled chillers to reject heat. These systems consume virtually zero water during normal operations.
When a facility does use water, it is frequently non-potable greywater or industrial reclaimed water that is unfit for human consumption anyway. Yet, the crowdsourced maps conflate a closed-loop facility with an open evaporative plant, generating mass outrage based on flawed engineering assumptions.
The Real Problem: The Local Economic Lie
If the environmental arguments are largely overblown by a lack of technical literacy, the tech industry’s defense mechanisms are equally dishonest. This is the critique the industry hates to admit.
For years, tech companies have sold local town councils on the lie of job creation. They walk into rural towns in Missouri, Texas, or Virginia promising a massive economic boom. They show slick renderings of pristine campuses and talk about the hundreds of jobs coming to the community.
The reality? A data center is an automated fortress.
During construction, yes, you get thousands of temporary union pipefitters, electricians, and concrete workers. But the day the facility goes live, the economic footprint plummets. A 100-megawatt facility can easily be operated by a skeleton crew of thirty people: a few security guards, some facility engineers to manage the HVAC, and a handful of technicians to swap out dead hard drives and server blades.
The high-paying AI research jobs, the software engineering roles, the data science positions—those people stay in San Francisco, Seattle, and New York. The hosting community gets a massive building that pays local property taxes but generates almost zero sustainable, long-term employment velocity for the local population.
This is the real point of tension. It is a real estate play, not an economic engine. When a community realizes they traded hundreds of acres of natural landscape for thirty night-watchman jobs and a slight bump in the municipal tax base, they feel cheated. That resentment is what fuels the sudden explosion of "anti-tech extremism" and local political revolts, not a genuine sudden passion for utility grid mechanics.
Dismantling the "People Also Ask" Delusions
To understand how deep the misinformation goes, you only have to look at the standard questions driving public discourse around this topic. The premises are fundamentally broken.
Do data centers cause widespread cancer or health issues?
No. This is where the Brockovich framing becomes genuinely counterproductive. By applying the same playbook used for chemical dumping scandals like the PG&E hexavalent chromium case, the public is led to believe that server farms are emitting toxic chemicals into the air or leaching carcinogens into the soil. A data center is full of silicon, copper, and aluminum. The primary emission is heat. The backup generators run on diesel, which requires standard industrial emissions permits and regular testing, but they operate only during rare emergency grid failures. Conflating digital infrastructure with industrial chemical manufacturing is unscientific fear-mongering.
Why can't we just move all data centers to the desert or Iceland?
Because physics doesn't care about your political preferences. Data centers are built where they are because of latency and fiber optic connectivity. If an AI model is running inference for an autonomous vehicle or a high-frequency trading platform in Chicago, you cannot host that model in rural Iceland. The time it takes for data to travel through fiber-optic cables across the ocean introduces a latency penalty that destroys real-time application performance. Furthermore, building in the middle of a desert complicates the very cooling and power issues activists are trying to solve. Facilities are built near major population centers because that is where the internet infrastructure already lives.
Stop Demanding Maps, Demand Capital
If you want to actually fix the friction between American communities and the infrastructure of the future, stop signing petitions for crowdsourced maps that tell you what any commercial real estate database already knows.
Transparency for the sake of transparency achieves nothing but localized NIMBYism. It shifts the infrastructure from one county to the neighboring county that is desperate enough for tax revenue to accept the bad deal.
Instead, local governments and community organizers must change their strategy from blanket opposition to aggressive, transactional capital extraction.
If a hyperscaler wants to build a multi-billion-dollar AI hub in your town, the answer should not be an outright ban or a emotional protest over noise. The answer should be a legally binding, ironclad community benefit agreement that forces the operator to directly mitigate their structural externalities.
- Mandate Microgrids: Require developers to build co-located, dedicated clean energy generation—such as on-site solar arrays or small modular reactors (SMRs)—that can back-feed the local municipal grid during peak demand events.
- Enforce Zero-Water Architecture: Ban evaporative cooling entirely at the municipal zoning level. Force operators to use dry-cooling or closed-loop liquid systems, regardless of the impact on their internal margins.
- Direct Infrastructure Taxation: Instead of accepting generic corporate tax breaks, implement specific municipal utility franchise fees that directly fund local public school districts, community colleges, and local water treatment plant upgrades.
The data center boom is not going to stop. You cannot run a 2026 global economy on localized 1990s industrial policy. The infrastructure will be built, either in your backyard or the next town over. Stop wasting energy trying to pierce a veil of corporate secrecy that is legally and physically mandated for national security. Accept the reality of the infrastructure, stop asking the wrong questions, and start forcing the richest companies in human history to pay the actual, un-subsidized cost of admission to your town.
