Initial Moratorium Proposal

Initial Ask Prior to Jan 6th Rezoning Meeting

Mar 05, 2026

I realized that I hadn’t shared the initial proposal and asked for a moratorium on the data center projects to facilitate increased research and critical questioning during the period between the Planning and Zoning Commission Meeting in Marana in early December and the Town Council Rezoning meeting on January 6th, 2026. Below is the full paper that was shared and presented to the town at the time to request a moratorium, addressing the first-pass issue set I had gathered in the initial research.

Baseline and Level-Setting on What Exists in Arizona

It is essential as part of this discussion, to understand the relative impact the proposed Marana projects would have compared to what currently exists in the state of Arizona. While exact figures are difficult to obtain, reliable public information indicates that 16 data centers consuming approximately 1,600 MW in the Phoenix area, with six facilities, the EdgeCore Mesa facility (450MW), Stream Data Center (280MW), Compass Data Center Goodyear (212MW), Aligned Phoenix (180MW), Vantage Data Center Maricopa (175MW) and Microsoft Goodyear (143MW) being the largest in size currently operating and the remaining being <20MW data centers⁽²⁾. Further estimates indicate approximately 81 completed and operating facilities, with 31 currently under construction and 126 planned, for a total planned consumption of 2,500 MW ⁽³⁾. Between these two data sources, we can surmise that, outside the six defined >100 MW data centers, the majority of what currently exists are data centers in the 30 MW-50 MW range.

Facility Size = Total Estimated Capacity / Active Facilities

42.3 MW average facility size = (1,600 MW + 2,500 MW) / (16 facilities + 81 facilities)

For comparison purposes, the TSMC Semiconductor Fabrication facility in Phoenix is a 200 MW facility, with a total of 1,000 MW when all six semiconductor fabs are completed and fully operational ⁽²¹⁾.

Power Infrastructure, Cost Ownership & Future Proofing

Let us perform a thought experiment to better understand the power impacts and ultimately the potential cost impacts. T

The two proposed 750MW facilities in Marana represent a fundamentally different facility from what currently exists in Arizona, specifically five times the average and 50% larger than the state’s current largest facility. With estimates of an additional 5,000 MW in planning within the Phoenix area, the 1,500 MW in Marana ⁽⁸⁾, 700 MW for Project Blue in Tucson ⁽⁷⁾, and 500+ MW for a potential facility on Davis-Monthan Air Base ⁽⁹⁾, overall power generation capacity is a critical concern. For further context, the peak daily demand for TEP, according to the Arizona Corporation Commission, in 2024, one of the worst summers on record in Arizona, was 2,917 MW ⁽¹²⁾, nearly the same as the proposed demand from the four project proposals noted above. Questions that have not been answered that are vital to understanding these proposals are:

What is the existing power generation capacity that the proposed utility partners can deploy?
What is the gap to the proposed needs represented by these facilities, as well as projected area growth over the comparable power-generating facility timeframe?
If there is a gap in required power-generating capacity, who is responsible for paying for the additional generation capacity needed to address it and ensuring a sustainable power infrastructure over the typical 30-50-year lifespan of generating facilities?
If there is a gap, where are the power-generating facilities planned for construction, and what are the potable and wastewater impacts of additional non-renewable generating facilities within these areas, and how does their potential consumption impact the water suppliers in the area?
What additional environmental and ecological impacts from these facilities can be expected, and what areas would be impacted?
What is the effect on the ratepayers for those utilities?

Renewable energy offset has been a critical element of the proposals for these facilities; however, it raises additional vital questions that are currently unresolved. With a typical density of 4-5 acres per 1 MW of solar power generation⁽¹⁰⁾, offsetting 500 MW of power from the proposed facilities would entail a solar farm of roughly 2,500 acres, or about 4 square miles. Even broken down, this represents multiple additional plots of similar size to the data center proposals themselves.

How does this fit into the overall Marana town plan and employment density projections for open space or public land?
What are the impacts on ratepayers for this infrastructure?
Are there any additional environmental or ecological impacts from the addition of such extensive facilities wherever they are located?

Additionally, for renewable sources, industrial-scale solar panels are manufactured in China (~80%), with Southeast Asia accounting for the majority of the remainder ⁽¹⁸⁾. Given the current tariff structures and import uncertainties, how do the availability and cost impacts of this approach affect the timing and viability of installing facilities of this size within the proposed projects’ constraints?

Without understanding these essential questions and how these critical power-generating factors align, not just with the proposed Marana projects but also with the area as a whole, since one of the core utility partners is the same, it is difficult to fully understand the project requirements. If facilities cannot be completed and brought online in time for deployment, how will the gap be addressed?

Backup Power Systems & Pollution Impact

While primary power generation and supply are critical, these facilities are required to operate 24/7/365 under defined uptime service agreements. Typically, hyperscale data centers meet Tier IV requirements ⁽¹¹⁾, which require 99.995% uptime, or less than 26 minutes of downtime per year. This level of uptime requires significant redundancy and reliability to ensure everything operates continuously, especially the facility’s overall power status. While TEP and Trico both boast high levels of grid reliability with 48 minutes ⁽¹³⁾ and 82 minutes ⁽¹⁴⁾ of downtime on average, respectively, in their latest reports, this is still significantly worse than the proposed facilities would require. To ensure the facilities meet their service-level agreements, backup power systems would be necessary, most likely diesel backup generators. With these generators, the question is again one of scale.

The emission regulation for Tier 4 Generators is 0.67 g/kWh for NOx emissions for industrial-scale backup power systems, based on the EPA Standard ⁽¹⁵⁾.

Generators for supporting facilities like the proposed typically have a maximum output of 4 MW per unit.

750 MW facility size divided by 4 MW ⁽²³⁾ per unit = 187.5 generators.

Adding a 25% factor of safety, redundancy, and parallelism for reliability, and rounding to the nearest whole number of generators, you get 235 diesel generators to support one facility.

The NOx output rate is calculated by:

Output Mass = Power * Emission Factor * Time

2,680g = 4,000 kW * 0.67 g/kWh * 1 hour

2.68 kg * 235 generators = 629.8 kg or ~1,400 lbs of NOx per hour of operation per facility.

Based on typical maintenance routines of running a generator for 30 minutes per month, per NFPA 110 ⁽²²⁾ to ensure it is operational, this results in 6 hours of operation per year, or 4+ tons of this pollutant entering the local environment. Given the prevailing winds in the region, which are South-Southeasterly to Southeasterly, according to the National Weather Service, people living in the Hardin - N. Wentz - N. Luckett - Marana Road region are squarely in the dispersion path of these pollutants. The essential question to answer for the community here is:

Given the volume of pollutants these large numbers of generators would produce, where does that go?
How does the thermal output of the facilities influence where the pollution goes? An independent third-party expert assessment of the distribution of pollution would be essential for understanding the health impacts in surrounding areas.
Additionally, with the power-generating facility proposed as part of the La Osa project west of Pichacho Peak in Pinal County and its associated facilities, they are in the process of reviewing where do the pollutants from those facilities go, and how does that affects the Marana projects and the area?
Furthermore, are there other alternatives that would mitigate the need for a diesel backup? If so, what are they and what are their impacts? If it is gas turbines, which are common in these applications, how does that affect the noise elements of the ordinance?
- While gas turbines can be better overall in terms of pollutant emissions, they have significantly worse noise emission performance, especially at very high and very low frequencies. How does the ordinance enable a compelling tradeoff to meet all of the needs for power supply while minimizing impact on neighboring areas?

Without a deeper understanding of the timing of closing power-generating gaps and the grid infrastructure upgrades required to safely and reliably transmit the power needed for these facilities, this pollutant estimate represents a best-case scenario. Additionally, a critical question for the utility companies is:

What is the readiness of the area power grids to not just carry, but sustain a disconnecting of a 750MW load on the system and not sustain widespread cascading failures or potentially catastrophic failures, when a load of that magnitude connects and disconnects from the grid?
How widespread would the impact be, and what would be the impact on all other rate payers with that kind of load spike propagating on the grid network? What are the implications of these spikes, especially in the summer when peak reliability is critical for survivability, given the extreme heat?

Compliance Mechanisms and Ensuring Efficacy

A well-known law in the technical world is Boehm’s Law, which states that errors are most frequent during requirements and design activities and that the more expensive they are to remove, the later they are removed. Numerous case studies have established that the cost curve for this correction is exponential, with the most expensive time to address issues in the requirements and design being when production is ongoing or completed. In this case, the town ordinance is the core requirement that must be met along with the facilities’ customers’ needs to ensure a successful deployment. Taking the time now to review, refine, and update the ordinance will help avoid costly changes and compliance efforts later, thereby mitigating future costs and complexities.

As an example of this with respect to the noise measurement components of the town ordinance:

“5. Before issuance of a certificate of occupancy or certificate of completion, whichever occurs first, the data center operator must conduct a post-construction noise study performed by a qualified third-party acoustic engineer to document sound levels emanating from the data center measured at the property line of the nearest property to the data center property that is planned or zoned for residential land uses, or other noise sensitive use as reasonably determined by the zoning administrator, during peak operation of the data center mechanical equipment. The post-construction noise study must demonstrate that sound levels do not exceed the maximum sound levels set forth in Table 2. If sound levels exceed the maximum sound levels set forth in Table 2, the town will issue a temporary certificate of occupancy until the data center operator demonstrates that the required sound mitigation is achieved. If the data center operator is unable to demonstrate compliance with the sound levels set forth in Table 2 before expiration of the temporary certificate of occupancy, the town will not issue a permanent certificate of occupancy.

6. The data center operator must also conduct an additional noise study, as measured at the property line of the nearest property to the data center property that is planned or zoned for residential land uses, or other noise sensitive use as reasonably determined by the zoning administrator, annually during peak operation of the data center mechanical equipment for five years after completion of the initial post-construction noise study and when requested by the town thereafter. The data center operator must provide the results of the noise study to the town within 30 days of the anniversary of the date on which the certificate of occupancy or certificate of completion was issued by the town. If sound levels exceed the maximum sound levels set forth in Table 2, the data center operator shall take appropriate steps to achieve the required sound mitigation.”

Within these two segments, the requirement for measurement is “during peak operation of the data center mechanical equipment”. Statements such as these are ambiguous in an engineering context and can lead to designs that don’t meet their intended use. In the case of this statement, what does “peak” mean? Is it based on the expected use of the facility, or the maximum use of the facility during the expected highest temperatures of the year? What internal temperature must be met within the facility to drive “peak” operation? Is it the minimum, maximum, or average? What happens if the test is done at one point in time and the summer is milder than usual, but the following year it is above average? Is the facility non-compliant? Improved overall clarity or specification of multiple scenarios within the ordinance will better help ensure precise alignment among the town, the developer, and the facility’s operators on expectations, leading to consistent compliance.

Enacting a moratorium and engaging with additional experts in the critical domains that facilities of this magnitude would impact will help the town ensure that it has the clearest, most transparent expectations for anyone who wishes to develop a facility of this nature in the town moving forward. Clarity and transparency will improve expectations, expedite approvals and reviews, and ensure that unintended consequences and missed expectations don’t lead to confusion and costly debates over compliance.

A further consideration for the town and its residents is a deeper understanding of the potential enforcement mechanisms and how to apply them to the scale of hyperscale developers and operators. Specifically with the two proposed Marana facilities, Beale Infrastructure’s parent company, Blue Owl Capital, advertises $3.2b in cash and undrawn debt capacity ⁽²⁰⁾. While a $2,500 per day per violation fine ⁽¹⁹⁾ is substantial for the current commercial, residential, or industrial clients, for companies with billions of dollars of accessible capital and further backers with tens of billions or more of cash on hand, these fines move from being a compelling enforcement mechanism to the cost of doing business. A deeper review and understanding of how to revise enforcement mechanisms to impact businesses of this size, so that enforcement drives compliance, is required.

Conclusion

A one-hundred and eighty (180) day moratorium to better understand these essential questions, as well as improve the town ordinance regarding data centers, is the best course of action to ensure a responsible, data-driven decision in the best interest of the residents of Marana. Substantial open questions about how this facility and the additional proposed facilities within the area impact each other and the ability of all stakeholders to meet the demands and ensure a solution that meets the expectations of all parties involved. A pause of this duration will also enable the town to engage further and look at alternative proposals and ideas for the growth and sustainability of Marana and its economy. The number of critical questions raised here with unknown but knowable answers is large enough that a moratorium for better due diligence is required to make the best possible decision for all parties involved.

References

Thanks for reading Sonoran Think Tank! Subscribe for free to receive new posts and support my work.

Sonoran Think Tank

Comments

Ready for more?