What the evidence shows

Concerns and questions, sorted into what’s documented, what depends on the project, and what’s not currently supported.

Data centers are a new kind of neighbor for most of Utah, and that newness has produced a wide range of claims — some well-supported, some still being studied, and some that aren’t backed by current evidence. This page sorts the most common concerns into three honest categories so residents and decision-makers can weigh them clearly.

We treat every concern listed below as worth raising. Sorting them is not the same as dismissing them. A concern in the third category may simply mean the science isn’t there yet, or that the claim has been confused with a different issue. Wherever possible, we link to primary sources so you can read for yourself.

● Well-documented impacts

Real, measurable, and consistent with the engineering of large data centers
The size of each impact depends on the specific facility, its cooling method, and how it generates power.

Large electricity demand
A single hyperscale campus can require hundreds of megawatts — sometimes more power than a small city. The U.S. Department of Energy and EPA both track this trend nationally.
Significant water use
Many traditional data centers use evaporative cooling, which consumes water. Closed-loop systems use much less. The EPA estimates U.S. data centers consumed roughly 17 billion gallons in 2023, with projections rising sharply through 2028.
Air emissions from on-site power generation
When a facility generates its own electricity from natural gas turbines, it produces nitrogen oxides, carbon dioxide, and other combustion byproducts. State air-quality permits regulate these emissions.
Noise from cooling systems and generators
Cooling fans, transformers, and backup generators produce continuous sound. Measured levels depend on equipment, setbacks, and barriers. Low-frequency noise can carry farther than higher frequencies.
Light at night
Operating 24/7, facilities can add significant outdoor lighting. This is manageable with shielding requirements but is a real change for dark-sky areas.
Construction-phase impacts
Heavy truck traffic, road wear, dust, and temporary noise are documented at every large project of this scale.
Electronic waste
Servers are replaced on roughly a three-to-five-year cycle, generating a steady stream of e-waste that must be recycled or disposed of properly.
Rate and infrastructure cost questions
Whether residential utility rates rise depends on how the cost of new generation and transmission is allocated between the data center and existing ratepayers. This is a policy question, not a fixed outcome — and it’s worth asking about every project.
● Concerns that depend on the specific project or site

Real possibilities, but outcomes vary widely
Depending on siting, design, regulation, and what was on the land before.

Property values
Published studies show mixed results. Proximity, visibility, noise, and the prior use of the land all matter. Some studies find small declines for immediately adjacent homes; others find no measurable effect at a distance.
Local respiratory health
If a facility generates its own power from gas turbines and is close to populated areas, increased NOx emissions can affect air quality. The size of the effect depends on stack height, emission controls, distance, and prevailing winds.
Groundwater and aquifer impact
Depends entirely on the water source. A facility drawing from a municipal system has a different impact than one drilling new wells into a shared aquifer. Closed-loop cooling reduces — but doesn’t eliminate — water use.
Wildlife and habitat
Site-specific. Lighting, fencing, fragmentation, and noise can affect bird and pollinator behavior. Mitigation depends on siting decisions and lighting design.
Fire risk
Lithium battery storage and diesel backup systems carry known fire risks. Risk to surrounding areas depends on the facility’s fire suppression systems, response plans, and the capacity of local fire departments.
Local job creation
Construction employment is large but temporary. Permanent operations staffing for a hyperscale facility is typically smaller than residents expect — often 50 to 200 people for a campus that cost billions to build. Indirect jobs (maintenance, security, contractors) add to that figure.
Ground vibration
Possible near large turbines or generators, but generally limited to areas very close to the equipment. Setback distances and foundation design largely determine whether it’s perceptible off-site.
● Claims not currently supported by evidence

Worth raising, but not backed by current research

These claims appear in public materials and online discussions but are not supported by current peer-reviewed research, regulatory findings, or established science. Including them here is not meant to dismiss the people raising them — it’s meant to be transparent about what the evidence does and doesn’t show.

Brain tumors, cancer clusters, infertility, bone density loss, or hearing loss caused by data centers
No epidemiological studies currently link data center operations to these health outcomes. Electromagnetic emissions from data center equipment fall well below regulated exposure limits, and the equipment is similar to that found in any large commercial building. If new research emerges, this page will be updated.
Livestock death, infertility in livestock, or chickens unable to lay eggs
These outcomes are not documented in connection with data centers in published research. Livestock health is affected by many factors (heat, feed, disease, water quality), and any specific complaints should be reported to the Utah Department of Agriculture and Food for investigation.
Loss of honeybee, butterfly, or songbird populations caused by data centers
These populations are under real pressure across the U.S. and globally, primarily from habitat loss, pesticide use, disease, and climate change. There is no current research attributing population decline to data center operations specifically.
Surveillance of local residents
Data centers store and process data for the companies that own the equipment inside them — cloud providers, AI companies, government agencies. They do not collect data on the surrounding community. Privacy concerns about how those tenant companies use data are legitimate, but they apply to the companies, not to the physical facility next door.
“Low-frequency cellular breakdown” or similar biological effects from facility emissions
This is not a recognized scientific phenomenon. Low-frequency noise can be annoying and, at high levels, cause sleep disturbance — which is a real and documented concern. But it does not cause cellular damage at the levels produced by data center equipment.

How to read this page

If a concern matters to you, the most useful next step is usually to ask a specific, answerable question of the developer or county — for example:

  • “What is the projected daily water use, and where does the water come from?”
  • “What are the modeled NOx emissions, and what’s the air-quality permit limit?”
  • “What is the projected nighttime noise level at the nearest residence?”
  • “Who pays for new transmission infrastructure?”
  • “What is the fire response plan, and has it been reviewed by the local fire district?”

Specific questions get specific answers. We’ve included a guide to public comment and project review on the For Residents page, and a printable list on the Resident Question Handout.

Scroll to Top