So, it’s election time, and data centers have become the latest political talking point.
But here are a few questions I haven’t heard anyone adequately answer.
Suppose a massive data center is built in your community. The developer pays for the buildings, roads, utility connections, and other infrastructure required for the project. In other words, the developer bears the upfront cost.
The data center then begins consuming enormous amounts of electricity and, depending on the facility, significant amounts of water. The data center pays its utility bills, of course.
But what happens to everyone else?
If this new demand requires additional electric generation, transmission lines, substations, water capacity, treatment facilities, or other long-term infrastructure, who ultimately pays for those costs?
Maybe your electric or water bill doesn't increase today or tomorrow because of the data center. But over time, if demand increases and billions of dollars of additional infrastructure must be built to meet that demand, doesn't some portion of that cost eventually find its way to other ratepayers?
And here is an even bigger question:
What happens if technology changes?
We are making enormous infrastructure investments today based on what data centers currently require. But technology moves incredibly fast. What if today's massive data center becomes outdated, substantially less valuable, or unnecessary 7, 10, or 15 years from now because computing becomes dramatically smaller, more efficient, and less resource-intensive?
What happens to all the infrastructure that was built to support it?
Who pays for the electric, water, road, and other infrastructure if the demand that justified those investments disappears?
That's the part of this discussion I think deserves much more attention.
I believe America needs data centers. We need the computing capacity to compete in AI and the technologies of the future.
But needing data centers doesn't necessarily mean we should concentrate massive numbers of them in one state or approve every project that comes along.
Maybe the smarter approach is more strategic: locate them where the infrastructure and resources can support them, spread the risk geographically, and, where practical, encourage development on existing industrial or commercial sites that are underutilized rather than consuming enormous amounts of undeveloped land.
I'm not against data centers.
I'm against making massive, long-term infrastructure bets without first answering a very simple question:
Who bears the risk if today's assumptions about demand, technology, and infrastructure turn out to be wrong?.