Colorado 2026 AI & Data Center Opinion Survey

Colorado 2026 AI & Data Center Opinion Survey Overview

Magellan Strategies is pleased to present the results of a survey of 1,389 registered voters in Colorado. The interviews were conducted from August 26 to September 2, 2026, using an MMS text-to-web methodology, and the survey has a margin of error of +/- 2.63% at the 95 percent confidence level. The survey data were weighted to reflect the voter registration demographics of the state, including gender, age, party affiliation, and congressional district.

The survey measured how Colorado voters view artificial intelligence and the data centers being built to run it. On AI, the questionnaire covered personal and workplace use of tools such as ChatGPT, Gemini, and Claude, overall favorability, the expected impact of AI on twelve areas of life, the effect of AI on jobs, comfort with local governments and school districts adopting AI, opinions of the companies and executives building AI, trust in AI companies on eleven measures, and the importance of staying ahead of Chinese AI companies.

On data centers, the survey measured familiarity, the information voters have seen, support or opposition to construction at four distances from home, the best path forward for local governments weighing a project, and agreement with eight statements about data centers. Five open-ended questions asked voters to explain their opinions in their own words, and the responses are summarized on this page and published in full in the linked verbatim documents.

This survey was designed and self-funded by Magellan Strategies. The questions were not developed with, shared with, or sponsored by any outside individual or organization prior to fieldwork. Magellan Strategies is a nonpartisan public opinion research firm, and the results are published to give local governments, school districts, and public policy organizations clear opinion data on two issues they are being asked to decide.

Any questions regarding this survey and its methodology can be directed to David Flaherty at [email protected] or 303-861-8585.

View and Download the Survey Results

Survey Topline Results Survey Presentation Survey Crosstabs Verbatim Responses to T5 Reasons for a Favorable Opinion of AI Verbatim Responses to T6 Reasons for an Unfavorable Opinion of AI Verbatim Responses to T21 Reasons for Supporting a Data Center in Their City or Town Verbatim Responses to T22 Reasons for Opposing a Data Center in Their City or Town Verbatim Responses to T23 Other Paths Forward for Data Center Projects

Key Findings: Colorado 2026 AI & Data Center Opinion Survey

A statewide survey of 1,389 Colorado registered voters, conducted August 26 to September 2, 2026, with a margin of error of +/- 2.63%. Data weighted to Colorado voter registration demographics. Funded by Magellan Strategies.

• Two-thirds of Colorado voters hold an unfavorable opinion of AI, and using it changes the picture: Sixty-six percent of voters have an unfavorable opinion of artificial intelligence, including 41% who are very unfavorable, while 29% are favorable. Sixty percent have used an AI tool in the past 12 months, and 43% of those users are favorable compared to 8% of non-users. Republican voters (48% favorable) and voters 65 and older (42%) are the warmest groups; voters 18 to 44 are the coldest at 16%.
• Jobs are the sharpest AI worry, and it cuts across every group: Seventy-seven percent of voters expect AI to eliminate jobs, with 52% saying it definitely will, and only 8% expect it to create jobs. Eighty-eight percent are concerned AI will eliminate jobs for future generations (59% very concerned), including 92% of voters 18 to 44 and 81% of voters 65 and older. Among the 897 employed respondents, 41% are concerned about losing their own job.
• Voters do not trust AI companies and do not want local governments or schools using AI: Seventy percent of voters are not comfortable with their local government incorporating AI into its operations, and 75% are not comfortable with their school district using it to educate students. No more than 16% trust AI companies to build tools that are accurate and reliable, and 83% say the tools are not safe for children. Meta draws an 86% unfavorable rating, and Google and Microsoft each sit near 66% to 68% unfavorable.
• Opposition to data center construction is broad, intense, and grows closer to home: Eighty-six percent of voters oppose a data center near their neighborhood, 81% strongly, and opposition stays at 82% in their city or town, 77% in their county, and 75% elsewhere in the state. Among the 1,162 voters who have seen information about data centers, 82% say it made them more negative. Republican voters are the most receptive group, yet only 34% support a data center in their city or town, and just 5% of voters 18 to 44 do.
• A majority want a ban on new data centers, and they reject the economic case for them: Fifty-five percent of voters say a ban on all data center construction projects is the best path for their community, 26% prefer a new zoning and land use code, and 4% favor a time delay. Support for a ban reaches 68% of voters 18 to 44, 63% of women, and 73% of voters who have not used AI. Eighty-two percent agree data centers use too much water, 78% agree they drive up household electricity costs, and 68% disagree that they bring long-term, high-paying jobs. Water in a drought state leads the reasons opponents give in their own words.

Use of Artificial Intelligence Tools

Sixty percent of Colorado voters have used an artificial intelligence tool such as ChatGPT, Gemini, or Claude in the past 12 months, and 40% have not. Use is higher among men (68%) than women (53%), and it is highest among voters 45 to 64 (69%) followed by voters 18 to 44 (61%) and voters 65 and older (47%). There is little difference by party: 62% of unaffiliated voters, 56% of Democratic voters, and 61% of Republican voters have used an AI tool. Income is the strongest predictor of use, rising from 54% of voters with household incomes under $75,000 to 82% of voters earning more than $200,000. Use ranges from 45% in the 3rd Congressional District to 72% in the 4th.

Among the 835 voters who have used an AI tool, 51% have used it for both personal and work tasks, 30% for a personal task only, and 15% for a business or work task only. Men (55%), voters 45 to 64 (59%), parents with a child in the household (63%), and voters earning more than $200,000 (77%) are the most likely to use AI in both parts of their lives.

Opinions of Artificial Intelligence

Colorado voters take a dim view of artificial intelligence. Sixty-six percent hold an unfavorable opinion of AI, including 41% who are very unfavorable, while 29% are favorable and 5% have no opinion. Only 9% of voters are very favorable.

Familiarity does not close the gap. Among voters who have used an AI tool in the past year, 43% are favorable, and even among those who use AI for both work and personal tasks, favorability reaches only 56%. Among voters who have not used AI, 8% are favorable. The sharpest divide is generational: Sixteen percent of voters 18 to 44 view AI favorably, compared with 40% of voters 45 to 64 and 42% of voters 65 and older. Men (38%) are more favorable than women (21%). Republican voters, at 48% favorable, are about twice as favorable as Democratic voters (20%) and unaffiliated voters (25%). Favorability also rises with income, from 20% of voters earning under $75,000 to 51% of voters earning more than $200,000, and it ranges from 23% in the 1st Congressional District to 40% in the 4th.

Reasons for a Favorable Opinion of Artificial Intelligence

The 401 respondents with a favorable opinion of AI were asked to describe their reasons in their own words. Five themes account for most of the responses.

Faster work and fewer mundane tasks. The most common reason is that AI gets work done faster. Respondents describe handing off repetitive or tedious tasks so they can focus on harder problems: drafting and cleaning up emails, letters, and reports; writing basic code and spreadsheets; taking clinical notes; and reviewing lengthy contracts for provisions that might work against them. Several call it an accelerator or force multiplier that turns weeks of work into an hour, and most frame it as a workplace tool first, with home use as a bonus.

A direct answer instead of a search. Many respondents value AI for quickly getting information in plain language, pitched to their level of understanding. A recurring comparison is with Google, which respondents describe as declining in quality and overrun by commercial interests. The uses are practical: photographing a broken appliance to learn how to fix it, troubleshooting a car problem, adjusting recipes, planning travel, getting gardening advice for local conditions, reading medical reports, checking whether a claim is true, and skipping the customer service line at Social Security, Medicare, and unemployment offices.

Running a business and working beyond one’s training. Small business owners and self-employed respondents say AI has enabled them to start or run a company they otherwise could not, compete with larger firms, and build their own software instead of paying for expensive subscriptions. Others describe coding without a degree, learning subjects beyond their formal education, preparing for job interviews, researching case law, or fact-checking a novel before publication. A few add that it is especially useful for people with learning disabilities or for anyone who could not otherwise afford an accountant, an attorney, or a second medical opinion.

The next industrial revolution and staying ahead of China. Some respondents focus less on their own use and more on where the technology is headed. They call it the new industrial revolution or the biggest change since the internet, say it is inevitable, and conclude that the sensible response is to learn it rather than resist it. Medicine comes up most often as the field with the largest payoff. A smaller group ties their view to national competitiveness, arguing that the United States must stay ahead of China and that AI will bring jobs and new business here.

Favorable with conditions attached. Many respondents qualify their favorable view before they finish the sentence. The most frequent condition is accuracy, since answers sometimes draw on unvetted or skewed sources and must be double-checked. Next come calls for regulation, guardrails, watermarking of AI-generated content, and liability for developers whose products cause harm. Data center water and power use draws objections even within this favorable group, along with job losses, students leaning on AI instead of learning, and artists’ work being used without permission. Several say they trust the tool more than the companies and politicians deploying it.

Reasons for an Unfavorable Opinion of Artificial Intelligence

The 918 respondents with an unfavorable opinion of AI were asked the same question. Their answers fall into seven themes, and the first of them is about data centers rather than the technology itself.

Water, power, and the environmental cost of data centers. A clear majority of unfavorable respondents start by describing what it takes to run AI, and water comes up more than anything else. They describe data centers cooling servers with drinking water while Colorado is in a drought, with active wildfire seasons and farmers and ranchers already short on supply. Electricity is the second complaint: strain on an already stressed grid, fossil fuel generators brought in to cover the new load, higher carbon emissions, and the open land and wildlife habitat that large campuses consume. Many call the trade absurd on its face, spending irreplaceable water and power so people can draft emails or generate pictures.

Outsourced thinking and declining skills. Close behind the environment, respondents believe the technology is reshaping how people think. They describe students turning in essays they did not write, teachers watching kids stop working to learn, a decline in literacy, and adults reaching for a chatbot for every question instead of reasoning it through. Several call it a crutch that lets the brain atrophy like an unused muscle. A smaller group extends the harm to mental health: chatbots used as therapists or friends by young or vulnerable users, systems built to agree with whoever is typing, and compulsive use.

Wrong answers and a product forced on everyone. A significant share do not trust what the tools produce. Respondents recount fabricated facts and sources, travel itineraries built around non-existent dates, and answers delivered so confidently that errors slip past people who take the output as fact. Customer service is a sore spot, with phone systems and chat agents that loop and never hand off to a person. Many object that the technology is being pushed on them anyway: AI answers inserted at the top of searches, AI features switched on by default in software they already own, and opt-outs buried or missing. Others worry about a public that can no longer tell generated images, video, and voices from real ones, and about deepfakes and voice-cloning scams aimed at older residents.

No rules and no trust in the people running it. Some of the opposition is aimed less at the software than at the companies and the pace. Respondents describe an industry moving faster than anyone can measure the consequences, with no meaningful regulation, no liability when the output is wrong, and executives who admit the risks and build anyway; they point to social media as the precedent for what self-policing produces. Related concerns include surveillance and data scraping without consent, military applications, and a speculative bubble that could leave communities holding empty buildings when it bursts.

Job loss for corporate gain. Employment is the central concern for many respondents. They say companies use the technology to cut staff rather than help existing workers, and that productivity gains go to shareholders instead of wages or shorter hours; some report being required to use tools at work that their employer is training to replace them. Entry-level and junior roles come up repeatedly, along with graphic design, writing and editing, customer service, and retail work. Respondents also note that a data center creates few permanent jobs once construction ends, and that no one has a plan for displaced workers.

Data centers as neighbors. The facilities themselves draw complaints from a meaningful minority: a round-the-clock hum heard by nearby residents and livestock, light pollution, heat vented into the surrounding area, and the loss of farmland and ranchland. Rising residential electric bills recur throughout, with respondents arguing that Xcel Energy and other utilities are passing grid upgrades for industrial customers onto households while tax abatements shift the public cost further onto residents. Several say local officials approve the projects over residents’ objections and that nobody got to vote.

Stolen work. A meaningful minority, including respondents who identify as artists, authors, illustrators, and musicians, object that the models were trained on copyrighted art, writing, music, and books without permission, credit, or payment, and that companies are now selling the results back while cutting the people whose work went in. They describe generated art and music flooding online markets and publishing with low-quality imitations that displace human work, and they object to the use of a person’s face, voice, and likeness without consent.

Opinions on the Impact of Artificial Intelligence

Voters were asked whether the impact of AI on twelve areas of life will be positive or negative. Medical and healthcare is the only area where a plurality expects a positive impact, at 40% positive and 31% negative. Every other area draws a negative plurality or majority. Voters are most pessimistic about the effect of AI on social media (74% negative), elections (73%), personal relationships (71%), how people get news and information (66%), and K-12 education (65%). About half expect a negative impact on how people do their jobs (49%) and on Colorado businesses (47%), with roughly a quarter expecting a positive impact on each.

Republican voters and voters 65 and older are the most likely to expect positive effects. On law enforcement, for example, 38% of Republican voters expect a positive impact compared with 13% of unaffiliated voters and 9% of Democratic voters, and 33% of voters 65 and older expect a positive impact compared with 7% of voters 18 to 44. Higher-income voters and voters who use AI for both work and personal tasks are also more optimistic across most areas.

Opinions on the Impact of Artificial Intelligence on the Job Market

Jobs are the center of the concern. Seventy-seven percent of voters expect the growth of AI to eliminate jobs, including 52% who say it definitely will and 25% who say it probably will. Eight percent expect AI to create jobs, and 13% expect it to create and eliminate jobs in equal measure. Men (12%), voters 65 and older (15%), and Republican voters (18%) are the most likely to expect job creation, but no group comes close to a majority. Expectations are similar in every congressional district, with 71% to 85% expecting job losses.

The worry is generational more than personal. Among the 897 employed respondents, 41% are concerned that AI will eliminate their own job, including 16% who are very concerned, while 50% are not concerned. Concern is highest among employed voters 18 to 44 (48%), Democratic voters (46%), unaffiliated voters (44%), and voters who use AI only for work (56%). By contrast, 88% of all voters are concerned that AI will eliminate jobs for future generations, and 59% are very concerned. That concern reaches 92% among voters 18 to 44, 91% among women, 95% among Democratic voters, and 95% among voters who have not used AI. Even among the least concerned groups, Republican voters (78%) and voters earning more than $200,000 (77%), more than three in four are concerned.

Opinions on the Use of Artificial Intelligence in Government and Schools

Voters do not want their public institutions adopting AI. Seventy percent of voters are not comfortable at all with their local government incorporating AI into its regular operations and the services it provides, while 25% are comfortable, including 6% who are very comfortable. Comfort rises with age, from 15% of voters 18 to 44 to 31% of voters 45 to 64 and 38% of voters 65 and older, and it rises with income, from 16% of voters earning under $75,000 to 49% of voters earning more than $200,000. Republican voters (36%) are more comfortable than unaffiliated voters (23%) and Democratic voters (20%). Voters who use AI for both work and personal tasks are the most comfortable group at 43%, compared with 10% of voters who have not used AI. Comfort does not reach a majority in any group measured, and it ranges from 21% to 33% across the eight congressional districts.

Voters are even less comfortable with AI in the classroom. Seventy-five percent are not comfortable at all with their local school district incorporating AI into how it educates students, and 20% are comfortable, including 4% who are very comfortable. Men (26%) are more comfortable than women (15%), and voters 65 and older (29%) are more comfortable than voters 18 to 44 (14%). The differences by party are small: 19% of unaffiliated voters, 20% of Democratic voters, and 26% of Republican voters are comfortable. Among voters who have not used AI, 8% are comfortable with school district use, compared with 34% of voters who use AI for both work and personal tasks.

Opinions of Artificial Intelligence Companies and Their Leaders

Voters were asked their opinion of eight companies that build, create, and manage AI tools. Among the companies most voters recognize, unfavorable ratings run from 66% for Google to 86% for Meta. Microsoft is viewed unfavorably by 68% of voters, SpaceX by 69%, Amazon by 70%, and OpenAI by 67%. Google (33%) and Microsoft (32%) have the highest favorable ratings. Two companies are unknown to a large share of voters: Thirty-six percent have never heard of Anthropic and 60% have never heard of Perplexity. Republican voters are more favorable toward every company tested, with the largest gaps on SpaceX (58% favorable among Republican voters, 8% among Democratic voters) and Amazon (55% and 17%).

Opinions of the executives leading these companies are lower still. Elon Musk is viewed favorably by 25% of voters and unfavorably by 74%, Jeff Bezos by 15% and 82%, and Mark Zuckerberg by 8% and 91%. Sam Altman of OpenAI is viewed unfavorably by 61%, with 29% never having heard of him, and Dario Amodei of Anthropic is unknown to 61% of voters. Musk is the most polarizing figure by party, at 65% favorable among Republican voters and 3% among Democratic voters.

Trust in Companies That Create Artificial Intelligence Tools

Voters were asked whether they trust companies to create AI tools that meet eleven standards. On no measure do more than 16% of voters say yes. The highest trust is for tools that are accurate and reliable (16% yes, 71% no) and safe in general (15% yes, 72% no). Eighty-three percent say the companies cannot be trusted to make tools that are safe for children or free from political bias, 82% say the same about respecting users’ privacy, and 80% say the companies cannot be trusted to be transparent, honest about their dangers, or accountable when something goes wrong.

Distrust is highest among voters 18 to 44 and Democratic voters. Ninety percent of voters 18 to 44 and 90% of Democratic voters do not trust the companies to make tools that are safe for children, compared with 75% of voters 65 and older and 70% of Republican voters. Voters who have not used AI are consistently less trusting than users, but even among voters who use AI for both work and personal tasks, 73% do not trust the companies on child safety and 54% do not trust them on accuracy.

Importance of Staying Ahead of Chinese AI Companies

Voters split on whether AI companies in the United States need to stay ahead of Chinese AI companies in developing the most advanced models. Forty-eight percent say it is important, including 26% who say very important, while 37% say it is not important at all and 15% have no opinion. This is the largest generational gap in the survey: Sixty-seven percent of voters 65 and older and 62% of voters 45 to 64 say it is important, compared with 30% of voters 18 to 44. Men (57%) place more importance on it than women (39%). By party, 70% of Republican voters say it is important, compared with 46% of unaffiliated voters and 34% of Democratic voters. Voters who use AI for both work and personal tasks (65%) and voters earning more than $200,000 (68%) are also well above the statewide figure. The 4th Congressional District (56%) is the only district where a majority says it is important.

Familiarity and Information Regarding Data Centers

Colorado voters know what data centers are and have been following the debate over them. Ninety-five percent are familiar with data centers, the large facilities that house computer servers, including 50% who are very familiar. Eighty-four percent have seen, read, or heard a lot (31%) or some (52%) information about the construction of data centers in Colorado, and 15% have seen no information at all. Awareness is similar across gender, age, and party, and it ranges from 78% in the 3rd and 8th Congressional Districts to 91% in the 5th.

What voters have seen has turned them against data centers. Among the 1,162 voters who have seen information, 82% say it gave them a more negative opinion of data centers, including 69% who say much more negative. Eight percent say the information made them more positive, and 7% say it was equally positive and negative. The share made more positive is 2% among voters 18 to 44, 19% among Republican voters, and 15% among voters who use AI for both work and personal tasks; in every group measured, at least two-thirds were made more negative.

Support and Opposition to Data Center Construction

Voters were told that data centers are large buildings that house computer servers used to store data and run online services, including artificial intelligence, and were then asked whether they would support or oppose the construction of a data center at four distances from home. Opposition is broad, intense, and grows as the facility comes closer. Seventy-five percent oppose a data center elsewhere in the state but not close to home, 77% oppose one in their county, 82% oppose one in their city or town, and 86% oppose one near their neighborhood. Strong opposition rises from 64% for a data center elsewhere in the state to 81% for one near the neighborhood. Support runs from 22% for a facility elsewhere in the state to 12% near the neighborhood.

Sixteen percent of voters support a data center in their city or town. Republican voters are the most receptive group at 34% support, compared with 12% of unaffiliated voters and 7% of Democratic voters. Support is 5% among voters 18 to 44, 24% among voters 45 to 64, and 26% among voters 65 and older. Among voters who have not used AI, 5% support a data center in their city or town, compared with 28% of voters who use AI for both work and personal tasks.

Reasons for Supporting a Data Center in Their City or Town

The 216 respondents who support a data center in their city or town were asked to explain why. Their reasons fall into five themes.

Jobs and tax revenue. Most respondents cite employment first, often listing nothing else. Those who elaborate separate the construction and trades work of the build phase from the technical and monitoring positions that follow, and describe the operating jobs as well paid by local standards. Tax revenue is the usual companion, with respondents expecting the added base to fund schools, roads, and county services, and some hoping it will lower homeowner property taxes or utility bills. A few say the construction work is temporary and the permanent headcount small, or that jobs are the only upside they can think of.

Inevitable infrastructure with room to build. Many supporters treat the facility as something that will exist somewhere, and it might as well be here. They describe data centers as the next utility, comparable to broadband, power, or telephone service, and note that the buildings predate the AI boom and already sit throughout Colorado. Respondents who use AI, cloud storage, or their phones for everything say it would be inconsistent to want the service and refuse the machines that run it. Others point to the open ground between the Front Range and Kansas, abandoned towns, and empty office campuses as places waiting for a use.

Keeping pace with China and neighboring states. Several respondents framed the facility as competition rather than local benefit. China came up most, with respondents describing AI capacity as a race the United States has to win and domestic data storage as a security question. A Colorado version of the same argument holds that the state is losing companies and talent to other states because of its business posture, and that turning away data centers would deepen that loss.

Water and power fears called overblown. Some respondents support data centers in part because they believe the case against them is flawed. They cite closed-loop systems, refrigerant chillers, and dry coolers to argue that Colorado facilities do not consume water as critics claim, and compare the figures in circulation to the water used to produce a hamburger or California almonds. Some fault news coverage and social media for the alarm, and a handful speak from experience in the industry.

Support with strings attached. Some support comes with conditions rather than as a blank endorsement. The most common condition is that the operator supplies its own power and water, pays its own way for infrastructure, and never pushes costs onto residential ratepayers; a few say their support ends the moment their utility bill rises. Siting comes next: on the outskirts, in industrial parks, in remote or underground locations, and away from neighborhoods, schools, ranches, and farms.

Reasons for Opposing a Data Center in Their City or Town

The 1,137 respondents who oppose a data center in their city or town were asked the same question. Water is the reason a clear majority give, and it is named more often than all other objections combined.

Water in a drought state. Most opponents say some version of the same thing: Colorado does not have the water to spare. Respondents describe the worst drought on record, a summer of wildfires, dead lawns under watering restrictions, wells and reservoirs running low, and the Colorado River drying up while downstream states still draw from it. Many argue that people, farms, and ranches should have first claim. A smaller group disputes the industry’s claims about closed-loop cooling, saying the systems still need water and are chosen because they are the cheapest option, and several say a facility that used gray water or no water at all would be a different conversation.

Electric rates and grid strain paid for by households. A significant share objects to the power draw and, more sharply, to who pays for it. Respondents expect their Xcel Energy or Black Hills Energy bills to rise to cover transmission upgrades and new generation for a single industrial customer. Others describe a grid already prone to outages, homes already asked to curb use during peak demand, and new load met with gas plants or diesel backup generators. Tax abatements draw the same complaint, and many say a company earning record profits should pay its own utilities and infrastructure without public assistance.

Noise, heat, light, and health next door. Respondents describe the constant sound most vividly, comparing it to a jet idling or a high-pitched whine heard for miles that never stops at night. People who moved to rural areas or small mountain towns for quiet say a data center would defeat the reason they came. Waste heat, around-the-clock lighting, construction and truck traffic, and falling home values follow, along with health concerns about infrasound, lost sleep, and effects on livestock and wildlife.

Farmland, open space, and Colorado’s scenery. A significant share object to what the buildings would replace and how they would look. Respondents describe windowless industrial boxes that do not belong in mountain communities, on the Eastern Plains, or on Weld County farmland. Wildlife displacement, habitat loss, and the conversion of working farms come up repeatedly, along with the view that the same land and water should go to housing, schools, parks, or a grocery store.

Few jobs, profits elsewhere, and a possible bubble. A meaningful minority reject the economic case on its own terms. Respondents estimate permanent staffing in the dozens at most, say construction crews are traveling specialists rather than local workers, and conclude that the headcount cannot justify the land, water, and power consumed. Some question whether the facilities will still be running in ten years and ask that developers post an escrow account to cover cleanup. Running alongside this is opposition to AI itself: respondents say they would fight a data center anywhere because they object to what it is built to do.

Distrust of developers, officials, and the absence of rules. Process and accountability are the core grievance for a meaningful minority. Respondents say companies understate water and power use, sign non-disclosure agreements with local governments, and abandon closed-loop and noise-mitigation promises as too expensive. Local officials receive the same treatment for approving projects opposed by residents and for failing to adopt ordinances on noise, light, water, and energy. Many note that Colorado has no statewide regulation of data centers and that no one voted on any of this.

Opinions on the Best Path Forward for Data Center Projects

Voters were told that many local governments in Colorado are deciding the best path forward for their communities regarding data center construction projects and were asked which option they prefer for their own community. Fifty-five percent choose a ban on all data center construction projects, 26% prefer developing a new zoning and land use code that specifically addresses data center applications, 4% prefer a time delay of 6, 9, or 12 months on all projects, and 11% offered another option in their own words.

The ban is the choice of 63% of women and 46% of men, and of 68% of voters 18 to 44, 47% of voters 45 to 64, and 38% of voters 65 and older. By party, 62% of Democratic voters and 58% of unaffiliated voters prefer a ban, compared with 38% of Republican voters. Support for a ban falls as income rises, from 65% of voters earning under $75,000 to 35% of voters earning more than $200,000, and it is 73% among voters who have not used AI compared with 36% among voters who use AI for both work and personal tasks. A ban is the plurality or majority choice in every congressional district, ranging from 43% in the 8th to 66% in the 3rd.

Other Paths Forward Suggested by Voters

The 11% of voters who chose some other option were asked to describe it. Four themes emerged.

Approve only if the facility pays its own way on water and power. The most common alternative is conditional approval: a data center can be built if it supplies or pays for its own electricity and water and leaves residential utility bills untouched. Respondents outline on-site solar, geothermal, or small nuclear generation; closed-loop or recycled wastewater cooling instead of fresh water; no draw on the Colorado River; and per-gallon pricing that rises with consumption. Several want local taxes on the operator, a performance deposit, or a contract that returns to the city and county for review every twelve months. Refusing tax breaks and incentives comes up repeatedly.

Slow it down with longer moratoriums, impact studies, and a public vote. Rather than a six- to twelve-month delay, these respondents want the pause to run longer or end only when something specific happens: a two-year or three-to-five-year moratorium, a halt until the drought is over, or a freeze while the state studies the effects of data centers elsewhere. Some would put each project to a yes-or-no vote by the people in the affected district, or require public hearings and an independent impact statement before any approval.

Keep them away from homes and water, or out of Colorado. Several respondents accept data centers only in certain locations, proposing a buffer from homes, schools, and parks; siting in remote areas with no aquifers; underground construction; reusing vacant office and tech buildings instead of farmland; and classifying data centers as heavy industry rather than light industrial. A smaller group said the state is the wrong place altogether.

Treat them like any other business. A number of respondents reject all three listed options because they want no data-center-specific rules. They would let the county planning board run its normal permitting process, allow construction in any zone that already permits industrial development, and impose nothing on data centers that is not imposed on a factory or warehouse. A few say Colorado should welcome data centers with incentives rather than restrictions.

Opinions on Statements About Data Centers

Voters accept the arguments against data centers and reject the arguments for them. Eighty-three percent agree that data centers put too much strain on the local power grid, including 69% who strongly agree. Eighty-two percent agree they use too much water (72% strongly), 78% agree they drive up electricity costs for nearby households (64% strongly), and 62% agree they are noisy, with 28% unsure on that statement. On the other side of the ledger, 62% disagree that data centers are essential for the future, 57% disagree that they bring in significant tax revenue for the local community, 68% disagree that they bring long-term, high-paying permanent jobs, and 77% disagree that the benefits to the local economy outweigh the impact on the community. Seventy-five percent disagree that local governments should offer tax incentives for data centers, including 68% who strongly disagree, while 12% agree.

Republican voters are the least critical group but still lean against data centers on most statements. Sixty-six percent of Republican voters agree data centers use too much water, compared with 85% of unaffiliated voters and 91% of Democratic voters, and 50% of Republican voters agree data centers are essential for the future, compared with 26% of unaffiliated voters and 21% of Democratic voters. Voters who use AI for both work and personal tasks split evenly on whether data centers are essential, at 50% agree, while 12% of voters who have not used AI agree.

Survey Demographics & Methodology