by Lynne M. Taylor, American Policy Center, ©2025
(Jul. 19, 2025) — It’s not new news that Data Centers are becoming quite an issue, all across America. These massive centers are also an issue, across the world. With the continued push for economic gains BEFORE people’s wellbeing, we’re going to see even more of these centers built.
This article seeks to point out some of the most ‘not so much talked about’ dangers.
Grow, Grow and Destroy:
Back in 2024[1], it was reported that a typical data center needs no less than 10 acres to lease or purchase, outright. The leases and purchases must involve homeowners or those who possess rural property (ie: farmers). Zoning will also need to be changed from residential to industrial zoning. Is your town or community capable of supporting such a shift? What types of conditions and legal requirements must be set up for compliance?
Also, in 2024, the State of Virginia was tagged as the location with the largest amount of data center growth in the world[2]. Think about that. How much of VA, from a historical perspective will be gobbled up? Is the goal of data centers to simply boost the economy? One must wonder.
Especially knowing that farming is at stake. So, we wipe out farms and history to erect buildings that house data tracking methods and use obscene amounts of natural resources for a profit margin?! VA is approaching 500 data centers, CA is second with just under 300. TX is third with just over 250 data centers. OH comes in fourth, with 150 data centers. IN rounds out the top 5 states with 147. Be sure you access the embedded resource to see where your state is, in the data matrix that’s becoming all too commonplace.
Considering that many of the Big Data companies are buying/leasing many more acres than the ‘typical 10’, what does this do to the water tables in the area? While many sources featured in this article will address water use, electricity costs. Many don’t readily bring up other issues that can cause destruction in a community or town.
What about the healthcare needs of those injured while data center construction goes on? Will the need for more critical care hospitals be a part of the Big Data companies’ plan? What about the town or community? Many of these data centers are being built in rural areas, where an emergency visit to a hospital is not always possible. While there are property services[3] available to see if your land is valuable to a data center, they can also offer several items to consider.
Many of those who are losing land are upset about it, not trying to profit from it. Think about it, why did Mr. and Mrs. Homeowner or Mr. and Mrs. Farmer choose their land? Because they saw value that’s priceless!
As of 2023, after the United States expected profit of almost $100 billion, these countries stand to make a ‘pretty penny’ are (in order) China ($65.5 billion), Japan ($17.3 billion), Germany ($15.2 billion), UK ($14.1 billion), France ($9.3 billion), India ($7.4 billion) and Canada ($7 billion). (Source: Statista Global Data).
So, how much do these same countries spend on data center healthcare related costs?
The US alone is seeing a $4.5 billion increase[4] in data center related healthcare issues that directly impact public health systems. Of this amount, how much is devoted to those who have been injured in construction and/or maintenance of the data centers? In 2002, the National Institutes of Health stated that $11.5 billion[5] went to construction related (non-fatal) injuries.
With the existing tariffs, it’s harder to pin down what the other countries from Statista Global Data’s information are spending on data center healthcare expenses.
Data Center Job Safety:
Recently, a vice president of a global communications organization was injured while helping build a data center. His harrowing account, as well as details of the extent of his injuries can be found on his LinkedIn[6] account. Here’s an excerpt, “On Friday 20 June while working at a new Data Center construction site, I was waiting to start some testing with some other vendors we all were standing well behind the safety barrier while another team was energizing the MV transformer loop. When they energized the third transformer, all three MV transformers failed and there were 3 large booms, then one of transformer oil tank peeled back forming a chute facing in the direction I was standing the oil came out like a volcano high in the air over the heads of people closer than me.”
This should raise some important questions; if injury can happen to an experienced data vp, what can we expect for those workers hired who haven’t been as seasoned about workplace dangers? This ties back to the previous point about healthcare costs and proximity to emergency services.
Read the rest here.
The peculiar claim is that we have to get “there” with generalized artificial intelligence first before the Chinese do. Yet, the limits on the Chinese getting the most advanced AI chips is slowing them down.
So, what is the prize? What does this AI dominance look like? We control the Chinese economy as a result? It is not clear from all the talk. I guess we have to assume we get all sorts of economic advantages. But I think these advantages have not been examined closely.
Look, we just found out that Microsoft has been using Chinese software folks for their Department of Defense contracts. There is no telling what damage has been done so far.
This is being sold as a proprietary battle between two competing software cultures. I need to spend a lot more time understanding what constitutes the cutting edge of these technologies. Meantime, I just remain skeptical of these panicked demands for more money and resources.
I question their confidence in the outcome. There are frequent reports about serious problems with large language models. They hallucinate a lot. They fabricate false information. They are are shown already to resist some instructions, act racist, and push users in dark directions. It’s crazy. These are not easy problems to solve. This race to build out these data centers while they still have these problems seems very unwise to me.