Where Colorado Data Center Developers Should Look Next

If you are acquiring land or planning a data center in Colorado, It’s time to take a closer look at how you define a good site. Cheap acreage, an industrial designation, and a receptive local government are good indicators, but that’s not holistic enough. I would also want a defensible water plan, room between the campus and residential growth, and a clear record of what each approval actually allows.

The recent developments in Weld County, Brighton, Arapahoe County, and Aurora have made my view of the eastern and northeastern Front Range more specific. I see a stronger case for existing industrial properties and carefully selected infrastructure corridors. I would be cautious about assuming that moving farther into rural areas will make a project easier to develop.

For data center companies, that means bringing the operating requirements into the land search earlier. For land developers, it means proving the site's suitability before marketing it as a data center opportunity. The strongest proposal should be able to explain why this property makes sense, how the facility will be served, and what the surrounding community can expect as it grows.

On September 9, Weld County commissioners voted 5-0 to approve Global AI's Use by Special Review permit for an existing industrial facility near Windsor. The Planning Commission had recommended approval 7-0. The county emphasized industrial reuse and the avoidance of agricultural land conversion. The approved first phase includes a closed-loop, air-cooled system, and future expansion requires further review. Those details are set out in the county's approval announcement.

That is a meaningful signal for developers considering industrial reuse. It gives you a concrete example of a county evaluating a data center through a defined land-use process. It also gives you a reason to read the approved scope carefully before assigning value to the rest of a property.

The political situation adds another consideration. Colorado Newsline reported on September 29 that two POWER petitions had been approved for circulation the previous day. One seeks a referendum on the Global AI approval. The other seeks voter approval before data centers receive occupancy permits. The report describes a 45-day signature period and a special election within 60 days if sufficient signatures are obtained.

I would keep Weld on the opportunity map. I would also separate the county's willingness to approve a project from the risk that organized opposition changes its path. Both belong in your underwriting. A unanimous vote gives you useful evidence about the application that was heard. Your acquisition schedule still needs to account for the public process around that decision.

I would also ask the team to identify which approvals remain outstanding. Weld explicitly says the land-use approval leaves other required permits and approvals to be obtained. If your financial model assumes the whole campus can proceed, every future phase needs its own documented basis.

Brighton's situation is still developing. Denver7's reporting on the proposed campus describes eight buildings of approximately 315,000 square feet each, or about 2.52 million square feet in total. The city had received a rezoning application, but the proposal had not reached a council public hearing. Mayor Pro Tem Peter Padilla said he intended to propose a 90-day moratorium at the October 6 council meeting.

As of October 5, I would record that as a proposed pause. Any acquisition or development decision should be updated against the action council actually takes. There is a substantial difference between anticipating a moratorium and having an adopted measure with defined terms, exceptions, and an effective date.

Padilla also acknowledged Brighton's need for development and industrial space in the Denver7 interview. I would read that as a reason to keep investigating the corridor while paying close attention to the standards the city may establish.

Water should be central to that work. Brighton's official drought guidance says Stage 1 measures have been extended through November 15, 2026. For a developer approaching the city now, I would want the water demand and proposed cooling design ready for scrutiny before asking anyone to support the project.

If a pause is adopted, use the time to resolve open questions. Ask how the proposed rules would apply to your parcel and application. Revisit your option deadlines. Give the owner a realistic account of the policy timetable. A corridor can remain interesting while the timing of an individual deal becomes harder to justify.

Arapahoe County is examining the industry while also working on its water planning. The county's September 22 data center study session materials describe a review of facility types, infrastructure needs, and regulatory approaches. Its proposed Water Supply Element brings existing water policies and recommendations from the 2024 Water Supply Study into a new comprehensive plan chapter.

For eastern Arapahoe County, I would put those processes side by side in the site review. The county's 2024 Water Supply Study provides the background for asking more useful questions about a property's supply. I would start with the actual provider and source serving the parcel, then work through the demand and development assumptions with that provider.

A simple water availability flag would tell me too little. I would want to know the volume available, the conditions attached to service, and how the proposed campus fits into the provider's long-term planning. I would ask what changes if the campus doubles in size or if its cooling requirements change.

Policy formation deserves continued attention from companies willing to do that work. You may have an opportunity to explain your design and help officials understand the requirements of a specific project. I would keep the commercial assumptions conservative until the applicable standards and service arrangements are clear.

That same discipline matters for land sellers. If you are presenting a parcel to an operator, include the current planning documents and explain the unresolved questions. A candid account of the site's status is more useful than a broad claim that the county is open for business.

Aurora adds a practical warning about the edge between industrial development and housing. Sentinel Colorado's September 28 reporting on the proposed Trident campus documents residents' objections to its location near their neighborhoods and concerns about the review process. Aurora's official drought page also identifies current Stage 1 conditions.

I would take those concerns into the site search before selecting a layout. Map occupied homes, approved housing, and land planned for future residential development. Ask where fans, generators, substations, loading areas, and security lighting would sit relative to those properties. Give the engineering team enough room to propose a workable arrangement.

For land developers, the question extends to the development you are planning around the campus. If you expect to sell nearby parcels for housing, consider how those future residents would experience the industrial use. I would want that relationship addressed in the master plan before either buyer commits.

An industrial zoning designation is an important starting point. I would still evaluate the neighborhood relationship independently and ask local counsel to explain the applicable review and notice requirements. Where residents have questions about a proposed facility, I would prefer to answer them while there is still room to adjust the design.

That approach may lead you to reject a parcel that looks attractive on paper. It may also identify an industrial corridor where the campus has more room to operate and expand. Either outcome improves the quality of the land decision.

The resource questions deserve precise answers. Lawrence Berkeley National Laboratory's 2024 United States Data Center Energy Usage Report estimates that U.S. data centers used 176 terawatt-hours of electricity in 2023, about 4.4 percent of the national total. It also estimates approximately 66 billion liters of direct water consumption that year. Those are national estimates across facilities with different designs. They provide context for public concern, but a local proposal needs its own demand calculations.

The report distinguishes direct water consumption at the facility from water consumption associated with electricity generation. It also examines cooling configurations with different energy and water requirements. I would ask your engineering team to explain those distinctions plainly and document the proposed design rather than relying on the phrase closed-loop cooling.

For your water submission, I would request annual demand, peak daily demand, initial filling requirements, and the assumptions behind each estimate. Identify any seasonal or supplemental cooling use. Explain the source of supply, wastewater arrangements, and what the operator proposes to do during drought restrictions. Put the first phase and full buildout in separate columns so readers can follow the growth plan.

For power, identify the facility load being requested, when each phase needs service, and the utility work on which those dates depend. Ask the utility to document the scope and conditions of its commitment. Have your team explain responsibility for upgrades and what happens if the project takes less power than expected.

Windsor's official Global AI project information page offers a useful example of why service language matters. The town describes a conditional water and wastewater will-serve letter tied to possible annexation and says no formal service agreement has been executed with the owners. I would read every service letter with that level of attention to conditions before relying on it in a transaction.

My preferred search would begin with existing industrial properties and rural or semi-rural infrastructure corridors. I would look for enough separation from residential growth, a credible route to utility service, and a cooling strategy suited to the available water supply. Agricultural conversion would need a clear justification in the site comparison.

Industrial reuse deserves investigation, especially where the property's history supports the proposed use. I would still commission the appropriate environmental and building diligence. Existing infrastructure should be evaluated for the actual operating plan, and any reuse assumptions should be supported by the relevant specialists.

For greenfield parcels, I would put a phase plan on the map early. Show the area needed for buildings, utility infrastructure, access, stormwater management, and operational buffers. Then test whether the remaining land can accommodate the expansion you intend to market. If it cannot, revise the advertised capacity before taking the site to operators.

I would keep a dated political and entitlement record for each candidate. Record the application, hearing, vote, petition, proposed rule, and adopted rule separately. Link the underlying document and identify the authority responsible for the next decision. Assign someone to update the record before major spending commitments.

That is the approach I would build into LandmAIn. A developer reviewing Weld should be able to see the approval and petition effort together. A Brighton prospect should show the proposed October 6 action until there is a verified decision. An Arapahoe parcel should carry the relevant water and data center policy documents. Each event should have a date and a source that the acquisition team can inspect.

For the transaction itself, ask counsel to align option periods and closing conditions with the approvals and utility milestones you actually need. Price the site against those requirements. If the owner wants the value of a fully serviceable campus, ask for the evidence that supports it.

I would bring the same specificity to public engagement. Explain the initial project, the possible expansion, and the decisions still ahead. Publish the demand estimates and their assumptions. Show how the layout addresses nearby homes and how the operator intends to respond if measured performance falls short of the commitments.

When presenting economic benefits, separate construction activity from ongoing employment. Identify which phase supports each estimate and have the tax analysis reviewed for the proposed incentives and infrastructure obligations. I would rather make a modest claim that can be checked than create expectations the operating campus may struggle to meet.

Some residents will remain opposed. Officials may ask for conditions that change the economics, and a company may decide to pursue another location. Those are outcomes your development process should allow for. Early engagement gives you a chance to understand the objections while the land and design decisions are still flexible.

My view of Colorado's opportunity is becoming more selective. I would continue investigating Weld, Brighton, eastern Arapahoe County, and Aurora with a clear account of each site's current status. The best candidates should combine a workable industrial location with documented service arrangements and a development plan the community can examine.

If you are bringing a site forward now, make that case before asking for the full commitment of capital. Show why the property fits the operation, how each phase will be supplied, and where the approvals stand. That is the level of preparation I would want behind a Colorado data center investment.