Data Cetner Discussion

Audience Questions and Responses

Matthew Gonzales of Chingon Strategies, LLC, presented a slide deck titled “AI, Data Centers, and New Mexico's Future: Opportunities, Challenges, and Community Considerations” to a public audience on August 10, 2026 at the Elk Club in Roswell, NM. Following this presentation, audience questions were transcribed, categorized, and answered below:

1. Zenith Volts Project

Questions submitted:

• How many residents in Chaves County need to sign to stop the Zenith Volts?
• The proposed project claims it will not use water, instead using geothermal-cooled water. Is this an existing technology? I'm worried about a first-time data center project using my home as their pioneer project. Is this relevant?
• What information is available for the data center proposed for Chaves County (Project Zenith)?
• If transparency is so important to success, why is there such a lack of information on the Zenith Volts project that has been permitted? Is Zenith Volts even a viable, reputable company? What determines their success if this is the first “foray into building data centers?”
• Is this a required public meeting
• No response from county commissioners. What is being hidden?
• Notice for people not on Facebook
• Phase 1 or 2 done
• Did you sign an NDA?
• Who do you work for? What is your official job title?
• Have you conducted an environmental impact study, and will that be available to the public?
• Note: I’m about 5 miles from the prospective site, and water pressure is already low

Response:

The August 10, 2026 session did not address any specific project proposed for Chaves County, including Zenith Volts (Project Zenith). This was by design, not evasion. The presenter, Matthew Gonzales, was not aware of the individual projects, their code names, or their viability, and had no role in permitting or evaluating them.

Several of the questions in this section assume the presenter was speaking on behalf of the city, the county, or a developer. He was not. Gonzales is the founder and CEO of Chingon Strategies LLC and a former mayor, school board member, and village administrator, an outside speaker with an energy and public-policy background rather than an agent of any project. He traveled to Roswell and presented at his own expense. Questions such as whether he signed an NDA, who he "works for," or what is "being hidden" therefore do not apply to this session; he had no project information to withhold, because he was never provided any.

It also follows that this was not a formal public hearing. It was an informational meeting that does not substitute for, nor satisfy, the statutory public-meeting requirements a specific project must meet. On that point, the presentation was clear about what does apply to every data center proposal in New Mexico: projects require air permits, environmental review, water review, public meetings, and regulatory approvals before they can proceed. In Gonzales's framing, "Communities need accurate information to make informed decisions," and the review process produces that information.

So for the project-specific questions, whether an environmental impact study has been done and will be public, what phase the project is in, and how residents not on Facebook are supposed to get notice, the honest answer is that this presentation is not the source. Those answers have to come from the county, the permitting agencies, and the developer itself, through the formal review process. There is a process in place. It is slow and thorough, but it is separate from the August 10 presentation.

2. Water Use, Sourcing, & Wastewater

Questions submitted:

Water Consumption & Capacity

• Why not build where there is an abundance of water?
• How much water does a data center use on average per year?
• What cost would using frack water add to initial project buildout?
• Explain the closed-loop system, and its positive or negative effect on water or wastewater.
• How much water will a data center require? Small unit, medium-sized unit, large units

Sourcing & Alternative Technologies

• What equity does responsible water consumption look like when they are already hurting for water?
• Produced water use: is that technology available currently?
• In a closed-loop system, explain the difference in water capacity compared to volumes to frack or horizontal drill oil.
• Are there currently any data centers using produced water? If so, what does that entail?
• What are the proposed water concerns and solutions for Roswell and the county given the rise of data centers and energy-heavy projects?

Impacts, Risks, & Enforcement

• We do not have water in southeast New Mexico; it is not Cimarron.
• Phoenix, Arizona, is now going to use saline water from Mexico
• How much water will be used
• Water contamination and its effect on neighbors
• If the data center corporation's actual water use exceeds the amount they have represented to the city or county, what enforcement tools exist? Will the city pursue legal action, impose penalties, restrict access, or renegotiate terms? Short, what is the city’s plan if actual water use does not match projected water use?

Response:

Both the presentation and the audience identified water as the single most important question a community should ask. Responsible water management appeared among the presentation's Keys to Success, and water consumption and land use were listed among the Common Concerns for Communities.

Over the years, cooling technology has evolved alongside data center growth, and the right approach depends heavily on local climate and geology. Cooling towers, which shed heat through evaporation, aren't a viable option in the arid Southwest and shouldn't be on the table here. Many of the country's largest data centers were deliberately sited where water is less of a constraint, but as geography becomes more important for latency and redundancy, new cooling systems have appeared in desert environments. One alternative, the closed-loop cooling system, uses a sealed loop to recirculate coolant instead of constantly consuming water. Closed-loop systems are also not all designed alike: some circulate proprietary chemical coolants, others use straight water, and the water footprint changes accordingly. Because of that variability, and because data centers can differ enormously in scope, no single answer fits every project. For comparison, Oracle estimates the closed-loop cooling system for Project Jupiter will need about 10 million gallons for its initial fill (~31 acre-feet), with roughly 168,000 gallons per year for ongoing maintenance. Geothermal cooling has real potential, but only where the underground geology supports it, with Deming cited as a positive example.

Produced water deserves its own note, because it comes up often and is widely misunderstood. In New Mexico, it is still treated as a hazardous byproduct rather than a ready resource. In May 2025, the Water Quality Control Commission banned any discharge of treated produced water to ground and surface water, a prohibition in effect through 2030; in May 2026, the Commission accepted an industry-led petition to begin rewriting that rule toward more permissive reuse. You can track that regulatory fight through the Western Environmental Law Center's coverage. For now, in-state reuse is confined to tightly controlled, no-discharge pilot projects overseen through the NMED produced water pilot program and the New Mexico Produced Water Research Consortium at NMSU.

As for whether produced water has ever actually cooled a data center: not yet at operational scale. The concept is being actively explored in the Texas Permian Basin, where oilfield wastewater is abundant, but regulators have not finalized a permitting path, and the extreme contamination of raw produced water runs up against the ultrapure water most cooling systems require. It remains a promising idea, but it is unproven in this application, and New Mexico trails Texas in both the law and the infrastructure needed to get there.

Underlying all of this is a point of New Mexico law that surprises many landowners: groundwater beneath your property is not privately owned. All water in the state is public property, administered by the New Mexico Office of the State Engineer under the doctrine of prior appropriation, "first in time, first in right." Buying or selling groundwater means transferring a state-issued permit, and unused rights can be forfeited. A commercial project cannot simply drill and pump; it has to acquire and transfer existing rights through the OSE. Crucially, only the consumptive-use portion of a right transfers. This is because agricultural irrigation returns a share of its water through the ground to the aquifer as return flow, slowly recharging the system. Thus, the state measures a right by what it actually consumes. Because industrial cooling returns little of what it draws, converting an agricultural right to industrial use reduces that right to the smaller consumptive-use figure. As a simplified illustration, an agricultural water right of 3 acre-feet might be reduced to 2.1 acre-feet of use for an industrial application.

Finally, on the Roswell-Artesian aquifer, the most relevant recent source is the January 2026 New Mexico 360 Groundwater Report, and it offers a qualified reassurance. The report is fundamentally a warning about statewide groundwater decline, but it repeatedly holds up the Pecos Valley Artesian Conservation District, the Roswell-Artesian Basin in Chaves and Eddy Counties, as one of New Mexico's genuine management success stories. It credits the district's strict metering order, enforceable water allotments, and a water-rights trading program with letting local users manage their aquifer while avoiding the priority calls and severe declines seen in places like the Ogallala aquifer near Clovis and Portales. The caution worth keeping is that this is a story about disciplined management, not limitless supply: the same basin saw substantial declines in the mid-20th century, and the report's core message is that no New Mexico aquifer should be treated as able to absorb open-ended new demand without metering and limits. Our basin is comparatively well-governed, which is a real advantage, but that resilience depends on the very metering and allotment discipline any large new industrial user would have to fit inside.
 

3. Environmental, Safety, & Wildlife Impacts

Questions submitted:

• Why put money toward something you know has a negative impact on our environment? We should not have to wager our well-being for unnecessary data centers
• What are the facts regarding pollution from site generators, noise pollution, and possible temperature increases, etc.? What are the plans to mitigate negative impacts?
• How many hyperscale data centers in the United States have been built on karst geology?
• What level of environmental review will be conducted?
• What happens to infrastructure when the project comes to an end? Recyclable?
• What is the long-term plan for when the GPUs are obsolete, say in 5 to 10 years? Possibly even as soon as 3 years? Will the company upgrade them or leave this site as is and abandon it?
• Who pays when the bubble bursts?
• What about sound and frequencies data centers generate?
• If an industrial accident or system failure at the data center results in contamination, what is the emergency response plan? Who bears the liability, and what protections exist for residents and agricultural users?
• Wildlife: migratory birds, horned lizards, antelope
• 250 acres of batteries that burn at 5,000 degrees. Safety. What is the plan in a disaster?
• Increases in heat
• What environmental impact studies
• Frequencies from data centers have been observed to sterilize herds of animals. Where are impacts on our farms/ranches and people?

Response:

Communities should absolutely be asking about air emissions, land use, traffic, noise, and light pollution. These concerns are real, and the presentation named them plainly. The reassuring part is that New Mexico's review is demanding, arguably more so than what a developer would face in neighboring states. Every project still has to clear air permits, environmental review, water review, public meetings, and regulatory approvals before it can proceed.

You do not have to take that on faith. Project Jupiter, the large Oracle and OpenAI campus proposed in Doña Ana County, drew more than 7,000 public comments on its air quality permits, and in response the state held a public hearing and extended its decision deadline. New Mexico's Air Quality Bureau also sizes a permit by calculating a facility's emissions as if it ran continuously, 8,760 hours per year. And the state has been tightening its air-permitting program rather than loosening it. That is the opposite of a rubber stamp. A comparable project in Texas or Arizona would generally face an easier environmental review path. Which leads to a key point of the presentation: if a data center is going to be built somewhere in the region regardless, it is better for our air, water, and wildlife that it be built here, under our safeguards, than across the line under weaker ones.

That said, facts should drive this, not fear. Some concerns are legitimate and belong squarely in the permit record: generator emissions, noise, heat, and effects on migratory birds, pronghorn, and other native wildlife. Other claims circulating locally, such as the notion that data center frequencies sterilize livestock herds, have no scientific basis and pull attention away from the measurable impacts the review process is built to catch. Communities deserve the real numbers, which is exactly what permitting is designed to produce.

Separately, several people raised remediation concerns: what happens when the technology moves on and the hardware, or the whole site, becomes obsolete? These are fair questions, and they split into two very different problems.

The first is the equipment. AI hardware turns over fast, with GPU-dense servers sometimes retired or redeployed after three years or less, but obsolete is not the same as waste. Retired accelerators frequently get a second life on less demanding workloads, and the hardware itself is dense with recoverable gold, silver, palladium, and rare earths. Handled through a certified recycler (R2v3 or e-Stewards), very little needs to reach a landfill; Microsoft, for one, has reported a reuse and recycling rate above 90% for its server components. The catch is that this happens reliably only when required and documented, because no standardized, industry-wide requirement currently forces data centers to prioritize reuse.

The second problem is the site itself, and this is where the oil-and-gas and wind comparisons become instructive. Those industries manage end-of-life through financial assurance: a bond or letter of credit posted up front so the operator, not the public, pays to plug, clean up, and restore the land when operations end. The lesson for any data center agreement is to require adequate, right-sized decommissioning security as a condition of approval. This ensures that if a developer walks away, the cleanup is already funded rather than handed to Chaves County taxpayers.

4. Economics, Tax Incentives, Jobs, & Funding

Questions submitted:

• If data centers are such good investments, why aren't they fully funded? Why do they need property tax forgiveness and LEDA funds? Property tax must be made up by other property owners, and taxes on wages paid by taxes are just returning the state's own money.
• 7/26/26 Forbes and Moody are reporting the top 5 biggest hyperscalers are racking up massive debt and projected to be IT by 2027. 
• What are communities/governments giving in tax breaks, etc? 
• Why put some of this funding into our educational system if you care so much about your children’s future? New Mexico does not have a board-accredited mortuary science program and Roswell does not have a cosmetology program. There are other opportunities you could be funding to make a better future for New Mexico citizens
• Will gross receipts taxes go up with the project?
• We have seen this before with the nuclear industry; they talk about economic opportunity, jobs, tech, and progressiveness in New Mexico. However, most jobs are filled by out-of-state contractors. What is the plan to be different and keep our money in New Mexico?
• How many local permanent jobs will be created?
• Will we consider something along the lines of a sovereign fund agreement with the center?
• Historical data center job creation is minimal
• Property value decline

Response:

The economic benefits of data centers can be overpromised and underdelivered, so the burden is on the community to negotiate real, enforceable value in exchange for whatever it gives up. The presentation's list of potential benefits is legitimate: new jobs, increased tax revenue, school funding, infrastructure improvements, workforce development, and economic diversification. So are its Keys to Success, which put community benefit agreements and long-term planning front and center. The leverage to secure those is real because these companies want sites with cheap, available power; a reliable grid; fiber connectivity; land; and water.

It helps to understand the economic incentives available in New Mexico. An industrial revenue bond is, in plain terms, a property tax and gross receipts tax subsidy. The company deeds the project to the county or municipality, and because property held by a county is exempt from taxation under the New Mexico Constitution, the project pays no property tax for as long as the bonds are outstanding, up to 30 years. Equipment bought with bond proceeds is also exempt from gross receipts and compensating tax. That is the "property tax forgiveness" people are asking about, and it is why residents' own gross receipts rate does not rise from IRB projects. Rather, the abatement flows to the company, not a new charge to neighbors. To partly offset the lost revenue, the county negotiates a payment in lieu of taxes, or PILOT. This is the critical detail for anyone worried about schools: for a data center, that payment is negotiated rather than fixed by statute, because only electric generation, transmission, and storage facilities are legally required to make payments in lieu of taxes to school districts. If school funding matters, it has to be purposely written into the deal.

These are not hypotheticals in New Mexico. Doña Ana County approved roughly $165 billion in industrial revenue bonds for the Oracle and OpenAI project, with likely significant 30-year property tax abatements. Years earlier, Los Lunas landed Facebook with up to $30 billion in bonds and a 100 percent property tax abatement over 30 years, against roughly 300 construction jobs and about 50 full-time positions in its initial phase (rising toward roughly 300 at full six-building buildout). That jobs ratio is the pattern, not an outlier.

The other common tool, LEDA, works differently. It is a discretionary "deal-closing" fund that reimburses a company for land, building, and infrastructure costs, requires enough projected economic return to offset the public investment, and requires collateral equal to the grant so the money can be clawed back if the company falls short. The clawback is the whole point, and it is where the efficacy question gets uncomfortable. A Legislative Finance Committee review of 101 LEDA agreements from 2016 to 2021 found that at least 33 failed to meet their job-creation requirements, and in about half of those cases the state either waived the clawback or did not pursue it. A tool is only as good as its enforcement.

On jobs specifically, the audience's instinct that "historical data center job creation is minimal" is well supported. Data centers are among the most capital-intensive and least labor-intensive facilities in the economy: a typical retail data center employs around 30 permanent workers, and even hyperscale campuses usually top out at a few hundred to at most a thousand. 

None of this means the benefits are illusory. Capital investment broadens the tax base, while wages for the permanent roles tend to run above local averages. Additionally, multiplier models credit every $1 million in data center payroll with an additional $1.2 to $1.8 million in local economic activity. Both IRBs and LEDA have their place in New Mexico's toolbox. The community's job is to treat the incentives as bargaining chips rather than gifts: right-size the PILOT so schools are funded, tie any LEDA dollars to milestones with clawbacks the state is willing to enforce, insist on a community benefit agreement, and ask directly how many of the jobs will go to New Mexicans. A sovereign-fund-style arrangement, which one attendee raised, is exactly the kind of term that belongs on that negotiating table.

5. Energy, Power Grid, & Infrastructure

Questions submitted:

• My concern is: where is the natural gas coming from?
• Will the power grid be used or affected in any way by this data center?
• AI centers vs worldwide. Data centers are not AI centers. SMR: small nuclear reactors. Have they been approved as power for AI?
• What redundant power do they consider?

Response:

Economic development at this scale runs on power, and New Mexico's defining feature here is a structural asymmetry. The 2019 Energy Transition Act sets escalating clean-energy mandates, requiring public utilities to reach 50% renewable energy by 2030, 80% by 2040, and 100% by 2045. However, a data center that generates its own electricity behind the meter is not a public utility, so these microgrid projects are not currently required to use cleaner energy sources and face less scrutiny from the Public Regulation Commission. A specific statutory hook also helps: Section 62-17-12(C) exempts energy generated by a qualifying microgrid of 20 megawatts or greater and sold to a utility from New Mexico's renewable portfolio standard until 2035. That is the "microgrid" advantage the presentation referenced, and it is why a developer can bring its own generation and operate largely apart from the grid.

As for what actually powers these projects, the ranking is right. Natural gas is the fastest path to production, which is why developers facing interconnection queues stretching five years or more are turning to behind-the-meter gas generation. Renewables carry their well-documented strengths and weaknesses and lean on battery storage to firm up 24/7 output. Nuclear in the form of small modular reactors remains years away: only two SMRs operate commercially anywhere in the world today, one in Russia since 2020 and one in China since 2023, and 2028 to 2030 is the earliest realistic window for the first U.S. units to power data centers, with broad availability more likely in the mid-2030s. So to answer the audience directly: no SMR is approved and operating as AI power in this country yet, and the 5-to-10-year horizon is a fair estimate.

That leaves the "where does the gas come from" question, which has a satisfying local answer. New Mexico sits on large gas reservoirs from the San Juan Basin around Farmington down through the Permian, and the Transwestern Pipeline Company's Roswell Compressor Station helps form the interstate backbone for gas transmission. In plain terms, transmission infrastructure on our doorstep is part of why gas-fired self-generation is so feasible here.

Finally, on redundant power: microgrids are typically designed with built-in redundancy and can either island from the grid or tie into it as backup, and whether the broader grid is affected at all comes down to that same behind-the-meter versus grid-tied choice. A self-contained project leans on its own reserve generation; a grid-connected one requires interconnection study and likely infrastructure upgrades (and PRC oversight), which is where cost and reliability questions for existing customers come back into play.

6. Governance, Policy, Public Process, & Transparency

Questions submitted:

• Would it be good if the state imposes an 18-month moratorium on any new data center development or construction?
• How many signatures are needed in New Mexico to stop data centers across New Mexico?
• When it comes to building and operating data centers responsibly, what are some examples of things we should look out for? Good and bad.
• What sources are available or examples of data centers that use the systems listed in the presentation?
• What data shows the US is behind in the data center race worldwide?
• What is being done to regulate who can build a data center?

Response:

This category gets to the heart of what the Zenith Volts questions were really about: process and trust. The presentation's Keys to Success led with local engagement and transparency for a reason, and the honest answer to most of these questions is that New Mexico's levers are real but are mostly local and legislative, not the kind of statewide up-or-down vote many residents assume exists.

On a state moratorium, this is a live debate, not a hypothetical. In July 2026 a group of Democratic legislators announced they will introduce a statewide moratorium on large-scale data centers in the 2027 session, which would pause new large-scale projects while the state builds a framework to evaluate water, energy, emissions, ratepayer, and community-benefit impacts.

On "how many signatures are needed to stop data centers across New Mexico," the answer is one people rarely like: no such mechanism exists. Apart from a narrow and difficult-to-obtain referendum process, New Mexico voters cannot petition a private project out of existence.

On regulating who can build, no single state data center license exists. Authority is spread across local zoning and permitting, NMED air and water permits, the Office of the State Engineer for water rights, and the PRC where a utility is involved.

Transparency deserves its own note, because it is where the Zenith Volts frustration and the presenter’s Keys to Success meet. New Mexico's real transparency tools are the Open Meetings Act and the Inspection of Public Records Act, and they have teeth.

Finally, on whether the US is "behind" in the data center race, by raw count it is not. The United States hosts more data centers than any other country by a wide margin. The competition framing in the presentation is really about maintaining leadership in AI and advanced computing against China, not about catching up on existing infrastructure. That distinction matters, because it reframes the local decision. The question for Chaves County is not whether America needs to race to build, but whether specific projects, on specific terms, are a good deal for the community.

Closing

Hard questions are exactly what this moment demands, and developers should have to earn the public's trust rather than assume it. At the same time, data centers are becoming part of the infrastructure that modern life depends on, so the real question is not whether they get built but how, and on whose terms. That answer is not dictated from a boardroom in another state; it is written locally, in the conditions a community is willing to negotiate and enforce. In the end, Chaves County will own the project it agrees to build, and the terms set at the front end are what determine whether that ownership becomes an asset or a liability.