Data Centers Are Coming: Who Pays the Power Bill?

I have opinions about a lot of things. Strong ones. And if you have followed me for more than five minutes, you already know I am not shy about sharing them.

Data centers are different.

Not because I am suddenly allergic to having an opinion, but because this is one of those issues where the vibe check is not enough. The details matter. A lot. And if we are going to do this in Colorado, and if we are going to do it in Weld County, we need to stop talking like it’s either the second coming of economic development or the death of civilization.

I recently spoke out positively about the one data center that has chosen Weld County as its home. Boy, did I get an earful – and not the usual “left vs. right” churn that typically dominates my inbox and my socials. This was from both sides of the aisle and from people who I consider pro-business and pro-growth. So it got my attention.

Here’s what I know, what I’m learning, and what I want from you.

First, the obvious. Data centers are coming. Not just to Colorado, but everywhere. The national conversation has moved from “How do we lure them?” to “How do we keep them from wrecking our power bills?” The Denver Post piece I’m using as the backbone for this lays it out plainly: states that used to roll out the red carpet are now putting their reading glasses on and asking who pays when a power-hungry industry shows up and asks for a small city’s worth of electricity.

And it is not paranoia. It is math.

The article points out that policymakers in multiple states are experimenting with new rate structures and requirements so regular people and small businesses do not get stuck underwriting massive grid upgrades. Oregon passed a law requiring utilities to charge data centers different rates than other customers because of how they can drive up costs for energy production and transmission. That is the core fear, cost shifting. If a developer needs a huge new substation and miles of transmission, and the bill gets smeared across everyone else, that is not “economic development.” That is a piñata full of taxpayer dollars, and we are the stick.

Now, let’s talk Colorado and Weld County, specifically.

Colorado regulators have been staring at this problem, too. Reporting in late 2025 highlighted a big issue: data centers can request far more power than they will ultimately use, which gums up planning and can push utilities toward expensive upgrades “just in case.” (The Colorado Sun) That same thread is showing up in industry reporting about interconnection backlogs at Public Service Company of Colorado as large-load requests pile up. (RTO Insider)

Meanwhile, Weld County is not pretending this is theoretical. The county has been actively considering and drafting data-center zoning language, including an ordinance to define data centers and specify where they are allowed. (Weld County) Local reporting has also pointed to specific interest and proposed projects near Windsor and the old Kodak site area, with developers like Global AI mentioned in the mix. (9News)

So, let’s do pros and cons like adults.

The Pros

Jobs and local spending, even if not a factory-style headcount.
Data centers are not massive employers once they are built, but construction is real money, and operations still require technicians, security, vendors, and ongoing maintenance. They also tend to pull in specialized contractors, which can be a small ecosystem boost.

Tax base and economic diversification.
This is the one that makes my conservative heart perk up. Weld is amazing at what we do, energy and agriculture have built a lot of families here, but it is smart to broaden the base. If a data center brings assessed value and stable tax revenue without chasing away existing industries, that is not nothing. That is resilience.

Private investment in infrastructure, if structured correctly.
This is the hinge. In the best version of this story, a developer comes in, pays for their own interconnection and system impacts, and does not socialize the costs. The Denver Post article describes states trying to force that exact outcome through long-term commitments and financial guarantees.

On-site generation and “bring your own power.”
This is where I get intrigued. Nationally, grid operators and regulators are exploring “bring-your-own-generation” concepts for large loads. (Reuters) If a data center wants to build generation, contract for it, or otherwise prove it can meet its appetite without raiding everyone else’s wallet, I see potential upside. That is not just consumption, that is.

The Cons

Electricity demand, grid strain, and the very real risk of cost shifting.
This is the monster in the room. The Denver Post piece references analysis showing wholesale electricity costs spiking dramatically in areas with significant data-center activity, while also acknowledging it is not always simple to assign blame. Still, the political response across states is telling: lawmakers do not trust the “don’t worry about it” assurances anymore.

Forecasting games and planning chaos.
One expert quoted in the article notes utilities can end up with wildly inaccurate forecasts because data center users shop around for sites. Translation: communities and utilities may gear up for a giant load that never materializes, and somebody still pays for the prep work unless we write the rules carefully.

Water, both direct and indirect.
I have said before that cooling technology is rapidly evolving, and I still believe that. But I am not going to hand-wave the concern away, especially in Colorado. Some facilities can use enormous volumes of water depending on cooling design, and even “low water” designs have an indirect water footprint through the electricity they consume. (IEEE Spectrum) If a developer says “minimal water,” the correct response is not applause. The correct response is, “Define minimal, show your design, show your WUE targets, and put it in writing.”

Land use, noise, and community fit.
These buildings are not cute. Fans, backup generators, substations, security lighting, truck traffic. It is not the apocalypse, but it is also not a winery. This is why Weld is looking at zoning definitions and where these uses belong. (Weld County) If we cram industrial-scale infrastructure next to homes and then act shocked when people get mad, that is on us.

Policy whiplash and incentive addiction.
States chased these projects with tax incentives for years, and now some are realizing they may have subsidized a problem. Colorado is already debating how hard to regulate, or require renewables, or structure tariffs. (Davis Graham) If the rules change midstream, communities can get stuck between “we want growth” and “why did my bill go up.”

So where does that leave me?

Honestly, still where I started, with a slight shift toward “cautiously supportive, aggressively picky.”

I am pro-property rights. I am pro-growth when it is well-managed. I am pro-investment. But I am also pro-ratepayer, and I do not want a scenario where regular families in Weld County end up paying more so a global company can run AI workloads 24/7 in a building the size of an aircraft carrier.

If you love data centers, cool. Bring your receipts. Show me examples where the tax base did not get crushed on utilities. If you hate them, also cool. Bring your receipts. Show me the bills, the water impacts, the noise issues, the busted promises.

Let’s do the rare thing: a good, old-fashioned debate without acting like toddlers in a sandbox.

Data centers. What do you think? (But don’t just vibe, cite sources so we can all learn.)