Weld County Data Centers: Set the Terms First

The next couple of Scott Sheets are coming to you from Washington, D.C. And yes, they might be a little shorter. Not because I ran out of opinions. Because I am at the National Association of Counties conference, and if you want to understand what local government is about to be forced to deal with, here’s the headline:

Data centers.

Not “data” like spreadsheets. “Centers” like a brand-new industrial use that shows up with a PowerPoint, a land broker, a lawyer, and an appetite for electricity that makes your average manufacturing plant look like a nightlight.

In every breakout session, from transportation to land use to human services, people are talking data centers. That is how you know it is real. Counties do not all spontaneously obsess over the same thing unless it is about to land in their lap with a zoning application attached.

And the politics of this are just getting started.

Here’s the part that the national narrative always misses. Data centers are not a red-state issue or a blue-state issue. They are a “your ratepayers, your water, your roads, your land use code, your election” issue. It is why you are seeing this show up in polling as something Americans can sense is coming, even if they cannot articulate a full opinion yet. A POLITICO poll found voters mostly expect data centers to become a campaign issue where they live, with nearly half saying it could happen within the next five years. And the public concerns are not abstract. People are worried about higher electricity bills, blackouts, and taxpayers getting stuck with costs.

Let me translate that out of poll language into county language: “Don’t raise my bill. Don’t break my grid. Don’t make me pay for your project.”

That is not hysteria. That is basic citizenship.

So let me say my part clearly, especially for folks back home in Weld County: I am supportive of data centers being sited here, cautiously supportive. If we do it right, it can be a win. If we do it sloppy, we will be holding the bag while somebody else holds the profit.

The whole debate boils down to siting and mitigation. It always does. Every time.

The two questions nobody gets to dodge

First: power.
Data centers are a load. A huge load. And the grid is not a vending machine where you push “AI” and electricity falls out.

Utilities across the country are openly telling investors they are increasing capital spending because data center demand is surging. Southern Company, for example, has talked about massive contracted “large load” demand and the need to spend big to serve it. (Reuters)

That spending is not imaginary. It is wires, substations, transformers, generation, transmission upgrades. Real steel and real cost.

Here is the rule that should be non-negotiable in every county, in every state, in every service territory:

If your project triggers grid upgrades, your project pays for them. Period.

Not “shared.” Not “socialized.” Not “spread across the rate base so nobody notices.” No.

And this is not just my conservative allergy to cost-shifting. It is a live regulatory fight right now, including at the federal level. FERC has an active proceeding on interconnection of large loads, specifically about doing this in a way that is timely and orderly and does not wreck reliability or fairness. (Federal Energy Regulatory Commission)

And the concern about cost-shifting is not theoretical. Groups like the Union of Concerned Scientists have warned about loopholes that can push billions in connection and upgrade costs onto customers instead of the large-load projects that caused them. (The Union of Concerned Scientists)

Even some of the tech world is starting to read the room. There are reports of companies publicly committing to cover grid upgrade costs tied to their AI data center growth, explicitly to avoid sticking households with the tab. (Business Insider)

Good. That should be standard practice, not a press release.

So in Weld County, if a data center wants to come here, they need to show up with commitments from the power provider that the project can be served without impacting availability of power or cost of power for existing customers. And if upgrades are required, those upgrades cannot be shoved into everybody’s bill like a hidden fee on a hotel receipt.

Second: water.
Cooling is changing. Some projects are moving toward different cooling methods, and that is a good sign. But “changing” does not mean “gone.”

In parts of the country, researchers and policy groups are documenting how significant water use can be for cooling, with numbers that range from “material” to “holy smokes.” (Environmental and Energy Study Institute)

So we do not get to do the politician thing where we say, “They’re high-tech, so I’m sure it’s fine.” No. We have to account for water demand up front, the same way we do for any major industrial use. What is the source, what is the consumption profile, what happens in drought, what happens when the project expands, and what is the mitigation plan.

This is where the governing complexity kicks in. Counties deal with land use. Utilities deal with power supply and cost recovery. States set the regulatory framework. The federal government sets pieces of the transmission and market rules. Everybody has a piece of the elephant, and the elephant is standing on our citizens’ electric bill.

And by the way, this is why data centers are becoming political. Because people can feel that the decision is happening “somewhere else,” and they are the ones who will live with it.

The upside is real, too

Now, before somebody clips this and says “Scott hates data centers,” relax. I like economic development. I also like not being lied to.

The revenue upside for local governments can be considerable. That is not hype. That is math. If you get the siting right, and if you get the agreements right, and if you avoid giving away the store in incentives, you can create a new stream of revenue that helps pay for real community needs.

Other states are wrestling with the incentive piece in public, including Virginia, where analysts have quantified the size of data center tax exemptions and lawmakers are now debating whether the public is actually getting what it was promised. (jlarc.virginia.gov)

That is another lesson for us: do not confuse “growth” with “good deals.” A deal can be large and still be bad.

In Weld County, we are staring down a reality where the state legislature has made it harder and harder to rely on the industry that carried Colorado’s economic weight for decades: oil and gas. I am not going to pretend that is not political. Democrats have pushed policies that functionally squeeze that revenue stream and the jobs tied to it. Weld County’s top revenue stream.

So yes, I see data center revenue as a potential stabilizer. A bridge. A replacement lane. But only if we treat it like grown-ups.

What now

If Weld County is going to welcome data centers, here is the framework I want us to keep repeating until it is boring:

Siting matters. Put them where impacts can be mitigated and infrastructure can support them.
No ratepayer harm. Power availability and power cost for existing customers are not allowed to be collateral damage.
No cost shifting. Grid upgrades required by the project are paid by the project.
Account for water. Cooling requirements are changing, but water planning is not optional.
Local benefit must be tangible. Jobs, tax revenue, infrastructure investment, and clear commitments that survive the next corporate reorg.

And then, the part I care about most: what we do with the revenue.

I have said it before, and I am going to keep saying it until somebody tells me to shut up or we actually do it: Weld should focus its resources and policies on fortifying the families who live here.

So if data centers become part of the next chapter of Weld County’s economy, I want to target that new tax revenue at the things that make a community resilient.

Infrastructure, including energy infrastructure. Roads. Public safety capacity that keeps up with growth.

And here’s an idea that should not be controversial but somehow always is: childcare.

If you want a pro-family county, you cannot just make speeches about family values and then act shocked when young families cannot afford to live here, or cannot find care, or cannot keep two incomes because the childcare math does not work. If we are serious about building a county where working families can thrive, childcare is infrastructure. Not the cute kind. The critical kind.

Data centers are not the devil. They are not salvation either. They are a tool, and like every tool, they can build something solid or they can take your finger off if you are careless.

We need receipts, agreements, mitigation, and backbone. If a company wants to plant a digital factory in Weld County, fine. But we are not doing it on vibes. We are doing it on terms that protect our people, our grid, and our future.