Pity the poor AI datacenters facing ‘discrimination’ | Arwa Mahdawi

The Physical Cost of the Cloud: Why AI Data Centers are the New Urban Frontier

For years, we’ve been told that the “Cloud” is an ethereal, weightless entity—a digital realm where our data floats invisibly. But the reality is far more grounded. The AI boom has transformed the cloud into something massive, loud and incredibly thirsty. We are witnessing a shift from “taco trucks on every corner” to data centers on every corner.

As artificial intelligence evolves from a novelty into the backbone of global industry, the physical infrastructure required to sustain it is colliding with the realities of land use, environmental limits, and the electric grid. This isn’t just a tech story; it’s a story about who pays the price for progress.

The Utility Tax: Who Actually Pays for AI?

One of the most overlooked trends in the AI gold rush is the “energy tax” shifted onto the average consumer. When a hyperscale data center plugs into a local grid, it doesn’t just use power—it changes the economics of the entire region. To accommodate this sudden surge, utilities often seek retail rate increases to fund infrastructure upgrades.

In the first half of 2025 alone, American utilities sought nearly $30 billion in retail rate increases. In some of the largest electric grids in the U.S., power prices have jumped significantly—some reports citing spikes as high as 76% in a single quarter—driven by the rampant demand of AI clusters.

Essentially, the retail customer is helping to foot the electric bill for the massive computing power required to generate AI images and LLM responses. This creates a parasitic relationship where the profit goes to Silicon Valley, but the cost is distributed among local homeowners.

Water Wars and Local Backlash

Beyond electricity, AI has a thirst problem. Data centers require millions of gallons of water for cooling to prevent servers from melting down. This is leading to direct conflicts with local communities over basic resources.

Take the case of Fayetteville, Georgia, where residents experienced sudden drops in water pressure only to discover a nearby data center had consumed 30 million gallons of water. In many instances, these facilities operate with tax incentives and discounted tariffs, further aggravating residents who feel their environment is being sacrificed for corporate gain.

The sentiment is shifting rapidly. Recent polling suggests that seven in 10 Americans now oppose the construction of AI data centers in their immediate vicinity. The “Not In My Backyard” (NIMBY) movement is no longer just about aesthetics; it’s about survival resources like water and grid stability.

The Legal Frontier: From Corporate Personhood to “Infrastructure Rights”

Perhaps the most ominous trend is the legal evolution of the data center. We have already seen the U.S. Supreme Court expand corporate personhood through rulings like Citizens United and Hobby Lobby, granting corporations rights to political speech and religious expression.

Now, this logic is being applied to the physical buildings themselves. In a recent clash in Ypsilanti Township, the University of Michigan faced a moratorium on water and sewer services for a proposed $1.2 billion AI data center. The university’s response? A legal threat claiming the moratorium was “unlawfully discriminatory” because it singled out data centers by label.

We are entering an era where the “rights” of a facility to exist and consume resources may be legally prioritized over the rights of human residents to have stable water pressure or affordable electricity. If a corporation is a person, it stands to reason that its “organs”—the data centers—might soon be granted similar protections.

The Energy Dilemma: Speed vs. Sustainability

The race to power AI is creating a fundamental conflict in energy policy. On one side, the grid needs speed. Solar and wind energy can be deployed quickly and are often the most cost-effective way to add new capacity to the interconnection queue.

On the other side, there is a push toward traditional “baseload” fuels—coal, natural gas, and nuclear. While these provide steady power, they are slower to build and often more environmentally damaging. This tension is leaving utility operators in a bind: do they prioritize the rapid growth of AI or the long-term sustainability of the planet?

If the industry continues to “stumble through” this growth, as OpenAI leadership has suggested, the result may be a fragmented grid prone to cascading outages, similar to the voltage fluctuations seen in Northern Virginia that recently disconnected dozens of facilities.

Common Questions About AI Infrastructure (FAQ)

Do AI data centers actually increase my electric bill?
Yes, indirectly. While the data center pays its own bill, the cost of upgrading the grid to support their massive load is often passed down to retail customers through rate hikes approved by utility commissions.

Why do AI data centers need so much water?
High-performance GPUs generate immense heat. To prevent hardware failure, data centers use cooling towers that evaporate millions of gallons of water to keep the equipment at operational temperatures.

What is “corporate personhood” in the context of AI?
It is the legal concept that a corporation has some of the same legal rights as a human being. In the AI era, this is being used to challenge local regulations that limit where data centers can be built or how many resources they can use.