AI Data Centers May Create More Uninhabitable Land Than Cell Towers

EMF, Power Quality, and the Electromagnetic Taking of Land

For the electrically sentient¹, how many cell towers would it take to render more than five million acres of land uninhabitable?

The modern blight of AI data center expansion is already raising concerns about noise, massive electrical demand, water consumption and pollution, excessive nighttime lighting, thermal emissions that may influence local weather patterns, and even infrasound.

The electromagnetic issues—not so much.

Let’s fix that.

Chief among my concerns are conducted emissions (dirty electricity) and AC magnetic fields.

In a recent discussion with Theodora Scarato, I asked a simple question: What is the area of impact when a data center is installed?

Her answer was striking:

“… Bloomberg found effects up to 50 miles, most within 20 miles.”

If Bloomberg’s reporting is accurate, a 50-mile radius represents roughly 7,850 square miles—or more than 5 million acres—of possible power-quality influence around major data-center activity.

The Questions Nobody Is Asking

And that only accounts for the reported concentration of highly distorted readings. If three-fourths were within 50 miles, what about the remaining fourth, how far does that extend?

More important: what can actually be done about it?

Will dirty-electricity filters, even thoughtfully chosen and implemented, meaningfully reduce these exposures?

Will they adequately address the relevant harmonics?

Are some portions of the disturbance carried through grounding pathways, or even the earth itself?

If so, can homes within such an area ever be brought down to exposure levels compatible with the long-term integrity of living systems—not merely the reliable operation of electrical equipment?

Would substantially increased electrical currents throughout the surrounding distribution system also increase ambient environmental magnetic fields?

If so, what would be their frequency content and what are the critical exposure levels for magnetic fields with frequencies other than 60Hz?

Who will monitor the monitors—making certain they are answering the right questions, the biologically relevant questions?

These the sorts of questions that should be resolved before AI infrastructure quietly continues the electromagnetic taking of land².

If we have largely failed to stop the reckless spread of RF radiation, how do we expect to confront emissions that are even less understood, less likely to be recognized by those approving the infrastructure?

How many county commissioners reviewing data-center proposals would even recognize these issues, would be capable of understanding?

How many utility planners?

How many consulting engineers would acknowledge that electromagnetic interference may warrant consideration not only for electrical equipment, but also for living systems?

Until those questions are taken seriously, we may continue evaluating the footprint of AI infrastructure in acres while overlooking its possible influence across vast landscapes.

Learning from History

History offers a cautionary lesson.

Large infrastructure projects are almost always introduced by emphasizing their benefits. Jobs. Economic growth. Tax revenue. Technological progress.

Those benefits are often real.

What history repeatedly shows, however, is that the full costs frequently become apparent much later—and are often borne disproportionately by those living closest to the project itself.

Railroads.

Dams.

Mining.

Oil and gas.

Transmission corridors.

Highways.

Again and again, the beneficiaries have often been regional or national, while many of the long-term consequences remained local.

The question is whether AI infrastructure will prove different.

Or whether, once again, we are evaluating today’s promises before fully understanding tomorrow’s costs.

Where Will We Live?

I think Roger Moller, a retired EMF surveyor from Britain, captured it well when he told me during our conversation, "You find a house, and I'll tell you if it's safe."

But where should one with electrical sentience even begin to look?

Finding lower-exposure places has become increasingly difficult even without considering AI data centers.

For years, many people simply moved to small towns where land was inexpensive, populations were lower, and RF exposures were often reduced.

That strategy may have already run its course.

The very characteristics that once attracted people seeking refuge—affordable land, available electrical service, and room for expansion—may also attract AI infrastructure and the workforce required to support it.

My preferred strategy has always been different.

Rather than simply seeking remote land, I have looked for lower population density combined with the effective use of terrain that naturally limits RF propagation.

Perhaps that strategy now becomes even more important.

If AI continues expanding as expected, the places most likely to remain relatively quiet may not simply be those that are remote, but those whose rugged terrain makes large-scale infrastructure and the supporting workforce more difficult, more costly, and therefore less attractive to develop.

One practical lesson has not changed: stay proactive.

It is far easier to search for a lower-exposure place to live while you are still healthy than after illness has already limited your options.

Also, advocacy is important, but it is also a long-term endeavor.

Like the familiar instruction on an airplane, put on your own oxygen mask before helping others.

Protect your own health first so that you remain able to help protect the health of others.

Conclusion

History has repeatedly shown that informed people make better decisions than uninformed ones.

AI data centers may prove to have an electromagnetic footprint that extends far beyond what most of us presently imagine. If so, the implications could extend well beyond human health to the broader biological systems that share our environment. Because the appetite for AI appears to be effectively unbounded, the expansion of the infrastructure supporting it may likewise prove effectively unbounded—making these questions increasingly important with each passing day.

My hope is twofold.

First, that those who are electrically sentient will remain proactive in finding lower-exposure places to live. It is far easier to do so before illness limits one’s options.

Second, that advocates, engineers, scientists, and policymakers will educate themselves on these questions and begin demanding answers before AI infrastructure quietly continues the electromagnetic taking of land—and before the remaining places of refuge become increasingly difficult to find.

Awareness precedes control.

Always.

If you’d like to explore these questions further, here’s my recent interview with Theodora Scarato, where we discuss the potential EMF implications of AI data centers.

Here’s a recent newsletter from Patricia Burke with articles on the topic as well.

You may or may not appreciate the first twenty minutes of contextual framing, but if you’re interested in the broader picture, I also recommend the long-form documentary by Truthstream Media. It provides an excellent overview of many of the non-EMF issues surrounding AI data center expansion.

I hope this is helpful.

___________________

¹ In a conversation I had with Andrew McAfee a few days before writing this essay, he used the term electrically sentient. I like it. Here I use it to describe individuals who perceive synthetic electromagnetic environments as biologically significant and potentially harmful.

² Historically, a taking meant losing ownership of land. Here I use the phrase differently. An electromagnetic taking occurs when land remains legally yours but no longer remains practically habitable for those seeking lower synthetic electromagnetic exposures.