I was driving home from work the other day when I found myself behind a Subaru. Naturally, it had one of those old-school “Coexist” bumper stickers on it, because apparently Colorado law requires at least one Subaru per county road to remind the rest of us how spiritually evolved its owner is while driving five miles under the speed limit.
The original point of that little rolling sermon was religious tolerance, or at least the aspiration that people with radically different beliefs could somehow live together without making every trip to the grocery store a holy war. But because my life is very glamorous, I was thinking about data centers when I rolled up behind this virtue-signaling Subaru driver. And the more I thought about it, the more that bumper sticker accidentally made sense.
Can Weld County coexist with data centers? I think we can. But coexistence cannot mean one side gets the tax base and the other side gets the noise, the water demand, the power strain, the road impacts, and a polite pat on the head from people who already live somewhere else. Coexistence only works when both sides have obligations. It requires rules. It requires honesty. It requires math.
That is why I was the lone “no” vote when Weld County approved its new data center code. I did not vote no because I oppose data centers. I do not. I am not afraid of technology, artificial intelligence, or the future, although I will admit some days social media makes a pretty good argument for going back to rotary phones and handwritten thank-you notes. I voted no because I did not believe the code put strict enough guardrails around facilities that could place enormous demands on Weld County’s water, power grid, land, roads, emergency services, and residents.
The county did take some important steps. According to Weld County’s own announcement, commissioners approved Code Ordinance 2026-01 on April 6 after roughly three hours of public comment, and the vote was 4-1, with me as the no vote. The ordinance requires an electricity “Will Serve” letter, proof of water, dBA noise regulations, a 65 dB(C) limit at the property boundary, and industrial zoning limitations. It also keeps data centers out of agricultural zones, which was an important improvement.
Those are not meaningless protections. But “better than nothing” is not the standard when we are talking about facilities that can behave like small cities plugged into the grid.
When I served on council and later as mayor in Johnstown, we had a simple policy for new subdivisions: bring your water. A developer did not just show up with a pretty plat map, a traffic study, and a consultant in expensive shoes. If they wanted to build homes, they had to bring the water needed to serve those homes. Usually, that meant water appurtenant to the land. What once watered crops would now water people. The developer would deed that water to the town, the town would treat it, and the future residents would receive service.
That was not anti-growth. That was growth with math.
The same principle should apply to data centers. If a data center wants to locate in Weld County, it should bring the amount of water needed to serve its facility to the water provider that will serve it. Not wishes. Not vibes. Not a glossy sustainability paragraph written by a communications department in another time zone. Actual water.
I understand and share the concern people have about water. Colorado water is finite. That is not a slogan. That is a fact of life here, especially when we are discussing these projects during a drought. People are right to ask hard questions. They are right to be skeptical. They are right to say, “Hold on, exactly where is this water coming from?”
They are also right to expect honesty. Newer data centers are not all the water pigs people have read about, and they are not all the monsters the professional outrage industry wants you to picture. Some newer closed-loop systems use a fraction of the water used by older evaporative cooling systems, and some designs are moving toward dry cooling. That matters. It matters so much that I wanted Weld County’s code to forbid evaporative cooling systems altogether. I was outvoted, which did not make me wrong. It just made me outvoted.
In a dry state, during dry years, with every gallon already spoken for by somebody, somewhere, allowing the most water-intensive cooling technology is not my idea of conservative planning. It is my idea of pretending the math will get bored and leave.
And while we are talking about water, let’s ask the impolite question: why can’t they bring it? I can already hear the objections warming up in the bullpen. Commissioner, that is expensive. Yes, and so are data centers. Commissioner, that is complicated. Yes, and so is building a warehouse full of computers that eats electricity like a small nation with a gaming addiction. Commissioner, the economics may not work. Fine. Then maybe the project does not work here.
For the last couple of years, I have watched with great angst as the City of Thornton moved forward with a massive pipeline to take water from farms they bought and dried up in my county and move it south to serve their thirsty municipality. Thornton describes the Thornton Water Project as a 70-mile underground pipeline carrying water from the WSSC reservoirs through Larimer, Weld, and Adams counties to Thornton customers, including 48 miles through Weld County.
I have lamented that project for a simple reason: when water leaves Weld County, opportunity leaves with it. That water once supported farms. Farms support families, equipment dealers, truckers, implement shops, grain elevators, feed stores, welders, mechanics, schools, churches, and towns. Water is not just water. In Weld County, water is economy. So if Thornton can build a pipeline to move water out of Weld County, why can’t Meta, Microsoft, Amazon, Google, or any other hyperscaler build one to bring water in?
Before somebody faints into a stack of engineering reports, I understand the economics are different. I understand water law is different. I understand rights, return flows, treatment, permitting, interstate compacts, environmental review, easements, pump stations, storage, and a hundred other things would make this hard. Good. Hard is not the same thing as impossible.
This country has built massive pipelines across states and regions to move oil and natural gas. Colonial Pipeline spans about 5,500 miles and describes itself as the largest refined-products pipeline by volume in the country. Kinder Morgan’s Cheyenne Plains Gas Pipeline consists of 410 miles of 36-inch pipe from near the Wyoming-Colorado border to south-central Kansas. The U.S. portion of the Keystone Pipeline included 1,084 miles of new, 30-inch pipeline across multiple states. The point is not that water is legally or economically identical to oil and gas. It is not. The point is that when America decides a resource matters enough, we figure out how to move it.
Well, water matters. And if money is truly no object for multinational corporations building the infrastructure of artificial intelligence, then let’s test that claim. Build the water infrastructure. Bring water from places that have too much to places that have too little. Build it big enough to serve your facility and create a lasting community benefit for the place you want to call home. Not a mural. Not a “community partnership” with tote bags. Water. Real, wet, useful water.
But, but, but, but, Commissioner, that would make it too expensive to develop in Northern Colorado.
Okay. Then develop somewhere else.
That is not hostility. That is honesty. Weld County should not be forced to choose between future technology and future water. If a project only works when local residents absorb the risk, local infrastructure absorbs the strain, and local agriculture absorbs the loss, then the project does not fit.
The same standard should apply to electricity. If municipal code can tell a residential developer, “Bring your own water,” then county code should be able to tell a hyperscale data center, “Bring your own power.” And I do not mean a polite little “Will Serve” letter from a power provider. I mean generation. Real generation. Dispatchable generation. Generation that can support the facility, tie into the grid, and strengthen the community instead of showing up like a 400-pound houseguest asking what’s for dinner.
Power generation is finite right now. Not because America forgot how to make electricity. We know how. We are very good at it when we are allowed to be. The problem is that building serious generation has become a bureaucratic pilgrimage through permitting, politics, lawsuits, ideology, and regulatory fog. By the time somebody gets approval to build a power plant, the original engineers are retired and the consultants have formed a support group.
Meanwhile, the grid is fragile. Demand is rising. Generation is strained. And Colorado has spent years telling utilities which energy sources are morally fashionable and which ones must be dragged behind the barn. Colorado’s HB19-1313 set targets for large utilities to reduce carbon dioxide emissions from electricity generation by 80 percent by 2030 compared with 2005 levels, with a goal of 100 percent reduction by 2050.
So no, I am not comforted by somebody telling me a massive data center can run on solar panels and good intentions. A hyperscale data center takes a serious amount of power. That means a serious amount of generation. And if the sun stops shining, the wind stops blowing, or the batteries run dry, you had better have something besides a press release.
This is where this Utility Dive piece gets interesting. The article discusses “Bring Your Own Power,” also called “Bring Your Own Generation,” as a model where large energy users supplement utility service with onsite or near-site generation. The point is not to replace the utility. The point is to work with the utility so a massive new load can come online without dumping grid-upgrade costs onto everyone else. The article specifically frames BYOP as a way to connect large loads faster, reduce strain on transmission infrastructure, and avoid sticking ratepayers with the bill.
That is exactly the kind of conversation Weld County should be having and codifying. Not “Can the grid absorb this?” Not “Can the utility write us a letter?” Not “Can we approve the land use and hope the electrons sort themselves out like responsible adults?” The question should be: what generation are you bringing with you?
That generation could take different forms. The Utility Dive piece focuses on fuel cells, which it says can provide reliable onsite or near-site power with lower local emissions, less noise, no water use during normal operation, and faster deployment than traditional infrastructure. That may be part of the answer, but the larger principle matters more than the sales brochure. If these hyperscalers want to build the infrastructure of artificial intelligence in Weld County, then they should help build the infrastructure of electricity in Weld County. Period.
Let the innovators innovate. Get government out of the way where government is the obstacle. Let serious companies build serious power. Let nuclear back into the conversation without everyone reaching for their emotional support clipboard. Let clean-burning natural gas do what it does, especially in Weld County, where we have plenty of it and know how to produce it. And yes, let renewables compete, too. I am not against solar. I am against pretending solar is a religion and physics is a hate crime.
If a data center developer wants to use onsite fuel cells, natural gas generation, advanced nuclear, batteries, renewables, or some hybrid system designed by people much smarter than the average legislative committee, great. Bring it forward. Show us the numbers. Show us the reliability. Show us how it ties into the grid. Show us how it protects existing ratepayers.
But do not show up in Weld County with a massive power appetite and expect everyone else to tighten their belts. Data centers should not just consume power. They should help create it. Used correctly, they could decentralize generation, make the grid more resilient, increase the total supply of power, and reduce the pressure that drives costs upward. Econ 101 still applies, even if it has been banned in several committee rooms.
So write it into code. Do not merely require a “Will Serve” letter. Require a power plan. Require generation. Require coordination with the utility. Require proof that the project will not make Weld County residents and businesses subsidize the data center’s appetite through higher rates, weaker reliability, or emergency infrastructure upgrades. Bring your own power. Better yet, bring your own generation.
But, but, but, but, Commissioner, that is hard.
Yup. So is developing artificial intelligence, and you fellas seemed to crack that.
But, but, but, but. Commissioner, that makes it too expensive to develop here.
Then maybe Northern Colorado is not ready for hyperscale data centers. Or maybe hyperscale data centers are not ready for Northern Colorado. Either way, that is not Weld County’s problem to solve by lowering the bar.
Noise and vibration are another concern, and people are right to raise it. A lot of what residents read online comes from older systems, poorly designed projects, inadequate setbacks, exposed cooling equipment, generators, transformers, and 24/7 mechanical hum that nobody wants in their backyard. Data centers can produce low-frequency noise that is more irritating than ordinary sound because it travels farther, penetrates barriers more easily, and may not be fully reflected by ordinary A-weighted decibel measurements. Larson Davis, a noise monitoring company, notes that low-frequency hum from data centers may not contribute much to A-weighted noise levels while still being the source of complaints, which is exactly why monitoring matters. (larsondavis.com)
That does not mean noise is impossible to control. It means Weld County code should require it to be controlled. Modern acoustic design can include site layout, acoustic barriers, silencers, louvers, enclosures, damping treatments, floating floors, resilient mounts, and vibration isolation for mechanical equipment. One technical paper on data center noise and vibration describes the use of resilient mounts and fasteners for equipment such as engine generators, pumps, fans, pipes, and ductwork, along with acoustic enclosures, barriers, and silencers. (acentech.com)
In plain English, don’t just bolt the giant humming machine to the floor and tell the neighbors to enjoy the future. Design it right, build it right, monitor it right, and if it violates the limits, fix it.
Weld County’s adopted ordinance does include both existing dBA noise standards and a C-scale limit of 65 dB(C) measured at the property boundary, which is important because dBC better captures some of the low-frequency sound that residents often describe as hum or rumble. But I do not think the county should stop there. We should require pre-construction acoustic modeling, post-construction verification, and continuous noise monitoring that is available to the public. (weld.gov)
Not buried in a binder. Not hidden behind a records request. Not summarized once a year by somebody named Chad from compliance. Public. Continuous. Understandable.
For years, I have worked to create predictability and certainty in code for business. That matters. Businesses should know the rules before they invest millions of dollars, and they should not have to guess what government wants this week based on which activist group yelled loudest before lunch. But Weld County residents deserve predictability and certainty too. They deserve to know the rules will protect them after the ribbon cutting. They deserve to know the county will not approve a project, cash the tax benefit, and leave nearby families listening to a low-frequency mechanical lullaby for the next twenty years.
Good code should protect both sides. The business gets clear rules. The residents get enforceable standards. Everybody gets the same scoreboard. That is how adults do land use.
I have received hundreds of emails about data centers. Many came from Weld County residents who are genuinely worried about water, power, noise, traffic, land use, and what these projects mean for the place they call home. I welcome that input. I may not agree with every concern or every proposed solution, but I respect people who live here and are trying to protect their homes, farms, families, and communities.
That is the conversation I want to have. What I do not welcome is the national outrage industry parachuting into Weld County with a clipboard, a donor list, and a pre-written script. Groups with national footprints have been organizing around data centers in Colorado. Moms Clean Air Force, for example, promoted a Colorado data center town hall for residents of Broomfield, Adams, Jefferson, and Weld Counties, calling for renewable power requirements, cost protections, and community safeguards. They are allowed to advocate. I am allowed to notice the pattern. (momscleanairforce.org)
When professional activists discover a new fight, they do not just bring concern. They bring fundraising emails, branding, talking points, and panic in a reusable tote bag. So yes, I will sit down with Weld County citizens. I will listen to residents who live near proposed projects. I will work with neighbors who want real standards, enforceable protections, and serious answers. But I am not interested in laundering the talking points of a Washington, D.C.-style 501(c)(3) outrage machine that treats Weld County like a stage set for its next donor campaign.
We have enough noise already.
Which brings me back to that Subaru. The bumper sticker said “Coexist.” Fine. Let’s coexist. But coexistence cannot mean the people already here absorb all the costs while the newcomers collect all the benefits. It cannot mean rural residents get the noise, farmers get priced out of water, ratepayers get stuck with grid upgrades, and multinational corporations get a ribbon cutting and a tax schedule.
That is not coexistence. That is colonization with better stationery.
Real coexistence means data centers fit into Weld County without hollowing out what made Weld County worth choosing in the first place. Bring your own water. Bring your own generation. Control your noise. Publish the data. Protect the grid. Respect the neighbors.
Then we can talk, because Weld County does not need to be afraid of the future. But we also do not need to be suckers for it.
Source: Utility Dive
