Data centers and regional impact
This guide translates complex municipal contracts and environmental data into a clear resource. Understand how industrial cooling functions in extreme heat: the reality of grid demand: and the legal gaps in public oversight. Every section contains verifiable sources.
El Paso context and the hyperscale challenge
Chihuahuan Desert
ClimateFinite aquifers
WaterIsolated grid
PowerIntegrating hyperscale data centers into municipal infrastructure involves more than a real estate transaction. This shift changes the daily water and electricity requirements for the region while placing waste heat and pollution near neighborhoods. El Paso features a Chihuahuan Desert climate, finite groundwater, and an isolated grid. A large facility places a heavy burden on these systems simultaneously.
Cooling is the primary engineering challenge in arid environments. When outdoor temperatures exceed 100°F (38°C), removing heat from a building requires significant energy or water. This efficiency loss is known as the desert penalty. Operators often choose between evaporative cooling or high electricity use for chillers. Residents deserve to see these trade-offs with clear numbers attached.
Regional data reveals a gap between corporate sustainability claims and local physics. Programs labeled as water positive often restore habitats far from the local aquifer. Renewable energy claims frequently rely on annual accounting rather than guaranteeing clean power during every hour of operation. This summary focuses on local questions: actual withdrawals, fuel consumption, and the enforcement of limits.
Cooling, permits, and local hydrology
2.5 million gallons per day
Permit scaleEvaporative loss
RiskWUE caps
Policy leverMarketing terms can obscure how cooling systems function. A closed loop often refers to the internal technology cooling system which circulates fluid over server chips. This internal circuit is sealed. However, the external facility loop rejects heat into the atmosphere through evaporation. Meta has secured a permit for up to 2.5 million gallons of water per day for the El Paso site. This permit confirms that the facility will evaporate large volumes of water during peak heat.
Corporate water positive goals often fund restoration projects in distant parts of a watershed. These credits do not physically return water to the local aquifer. Neighbors in other regions have noted that water returned to a general watershed is different from water available for local drinking.
Artificial intelligence workloads increase thermal density, requiring 50kW to 100kW per rack compared to 5kW for traditional servers. This shift drives up cooling demand. Effective policies tie limits to intensity metrics like liters per kilowatt-hour. Communities should require reclaimed water for industrial cooling to protect potable supplies.
- Request peak daily withdrawal totals and drought rules.
- Verify if makeup water is potable or reclaimed.
- Require binding intensity limits over one-time press releases.
- Establish a Water Usage Effectiveness (WUE) limit below 0.3 L/kWh.
Crisis Priority and the 'Equality' Trap
Section 14.9
Legal HurdleIndustrial parity
PriorityPublic health
TriggerSection 14.9 of the 380 agreement creates a major hurdle for emergency resource management. The City is prohibited from enacting new policies that specifically target the project's water or power consumption. Any conservation mandate must be 'generally applicable on a City-wide basis' to remain valid. This clause ensures that the data center maintains the same resource priority as a local hospital or a family home during a shortage.
This creates a situation where the City cannot prioritize households over industrial cooling during a severe drought. If the aquifers drop to critical levels, the City must apply restrictions to all residents before it can legally force the data center to reduce its intake. The agreement protects the facility from being singled out for its massive consumption patterns.
Public health and safety is the only legal exception to this rule. However, the 'generally applicable' requirement remains a high bar for local government. This prevents the City from using its police power to manage the data center as a unique industrial risk. Residents should ask if this parity is appropriate for a facility that evaporates millions of gallons per day.
Treating a hyperscale facility as a standard customer during a crisis places a heavy burden on the community.
- Drought Lock-in: The City cannot pass 'data center only' water cuts, even if the facility's 2.5 MGD use threatens the local water table.
- Grid Stress: During peak heat events, the City cannot legally demand the facility power down unless they mandate similar cuts for the general public.
- Legal Litigation: Any attempt to prioritize residential needs over the facility could trigger a breach of contract and a lawsuit from the parent company.
Grid load, modular gas, and hourly reality
366 MW
McCloud Gas PlantYears
Bridge periodSanta Teresa
Regional clusterEl Paso Electric proposes the McCloud natural gas plant to power the Meta data center. This 366 MW facility would consist of 813 modular natural gas generators. A proposed bridge strategy allows the customer to fund the asset for five years before it integrates into the general ratepayer base. Open proceedings are necessary to weigh the long-term impact on bills and climate goals.
Renewable energy pledges usually match production annually. The grid requires balance every second. Without large-scale storage, facilities run on fossil fuels at night when solar generation is unavailable. Hourly matching ensures green power is available at the specific time of use.
Regional grid strain is worsened by nearby developments like Project Jupiter in Santa Teresa, New Mexico. Clustered facilities create a resource sink. Cumulative planning is essential because individual statements often overlook the combined stress on transmission systems.
- Review utility dockets for details on infrastructure costs.
- Ask who covers costs for new plants after the initial bridge period.
- Verify if renewable claims account for nighttime AI training loads.
Ozone, backup power, and point-source pollution
Non-attainment
Ozone context366 MW
McCloud (proposed)Demand response
Watch itemHyperscale sites rely on large backup diesel and gas-fired systems so power stays available during tests and grid stress. El Paso sits in a federal ozone non-attainment basin. Ozone forms when nitrogen oxides and other compounds bake in sunlight and heat.
El Paso ranked ninth among major U.S. cities for annual average fine particulate pollution (PM2.5), at 7.8 micrograms per cubic meter. The same IQAir release, ranked Los Angeles highest among major U.S. metros for PM2.5, with San Antonio, Houston, and Fort Worth above El Paso among Texas cities (worse particulate levels). An IQAir science manager told reporters that Texas has a relatively sparse air quality monitoring network for its size, so not every neighborhood has hyper-local readings.
That El Paso Matters piece also described early 2025 dust storms that yielded four of the five most polluted days on record at EPA's long-running Chamizal monitor in South Central El Paso (daily data since 1999) and two of the worst days ever at a monitor in Socorro in far East El Paso County. The story tied the dust to record heat, long-running drought, and stronger-than-average spring winds kicking up bare soil. Those swings show why annual rankings alone understate bad-air weeks when new industrial sources are proposed.
On-site equipment can create concentrated air quality risks next to the fence line.
- Diesel particulate matter (DPM): Large diesel generators used for emergency power and routine testing emit fine particulate matter (PM2.5) that can penetrate deep into the lungs. That stacks on top of regional PM2.5 pressure in a city that national surveys already rank among the worst for big-city particulate averages.
- NOx, carbon monoxide, and smog: When diesel or gas turbines run, they release nitrogen oxides (NOx), a primary ingredient in ground-level ozone. Hot, sunny El Paso days already convert NOx and other precursors into smog in the local airshed.
- Health near point sources: Nearby residents can face higher asthma rates, reduced lung function, and cardiovascular stress from repeated exposure to plumes from backup testing and operations.
Steady megawatt demand often comes with new power plants and fuel delivery networks.
- McCloud natural gas plant: El Paso Electric has advanced a proposed 366-megawatt natural gas plant (McCloud) in public filings tied to serving the region's data center load.
- Smog and ozone: A new gas plant adds NOx and other ozone precursors in a basin that already struggles with federal ozone standards. Stagnant summer heat can make secondary smog formation worse.
- Methane along the gas system: Pipelines, compressors, and other gas infrastructure can leak methane, a potent greenhouse gas that worsens long-term climate instability.
Desert conditions amplify how industrial and backup emissions mix with other pollution.
- Urban heat: Data centers vent enormous waste heat from servers and cooling systems. In El Paso's desert climate that can strengthen local heat-island effects, keeping neighborhoods hotter than surrounding open land.
- Dust and particles: Press coverage of IQAir's 2025 assessments tied a sharp rise in measured pollution in El Paso to historic dust storms, including figures on the order of a 46 percent increase summarized in national reporting. Construction dust and industrial plumes can also mix with windblown dust, worsening irritating particulate loads for people breathing outdoors.
Comparative research in other dense U.S. data center markets has linked large backup generator fleets to regional pollution burdens similar in scale to conventional power plants, a benchmark El Paso should use when sizing test hours and stack limits.
Modular natural gas generators emit significantly less particulate matter than traditional diesel units. However, gas infrastructure creates a permanent fossil fuel commitment. Some programs pay operators to run these generators during peak grid stress rather than true emergencies.
Transparency should include the frequency and duration of backup operations. Agreements should prohibit running generators for economic profit through demand response programs.
Urban heat, the desert, and waste heat from digital infrastructure
Desert + UHI
Regional pattern~3.6°F LST
Data heat island (study)6.2 mi
Spatial footprintThe urban heat island (UHI) effect occurs when cities replace natural land cover with dense concentrations of pavement, buildings, and other surfaces that absorb and retain heat. In a high-desert environment like El Paso, that pattern creates specific challenges for public health and infrastructure, including added pressure on ozone chemistry and the electricity grid when temperatures stay elevated.
National analyses rank El Paso among U.S. cities with large measured gaps between urban heat intensity and tree-supported cooling, which matters for where new industrial heat sources are sited. Community heat mapping in the region also documents sharp block-by-block differences in temperature and heat index.
A March 2026 preprint by Marinoni et al. uses decades of land surface temperature (LST) from remote sensing to isolate what they call the data heat island (DHI) effect around AI data centers worldwide. They estimate that after operations begin, surrounding land surface temperatures rise by about 3.6°F on average (about 2°C), with extreme uplifts above 16°F in arid settings. In the largest cases they report LST surges up to about 16.4°F (9.1°C) near major facilities. The warming footprint is not fence-limited: increased LST is detected out to about 6.2 miles (10 km) from sites. The authors estimate that more than 340 million people could be affected by this temperature increase worldwide.
Built surfaces and equipment change how solar energy moves through the city.
- Thermal mass: Materials like asphalt, steel, and concrete have high heat capacity. They absorb solar radiation during the day and release it slowly at night, which can prevent the city from cooling down as much as the surrounding desert.
- Waste heat: Air conditioning units, vehicle engines, and industrial facilities (including data centers) vent waste heat directly into the local atmosphere, further raising ambient temperatures.
- Evapotranspiration: Trees and plants cool the air by releasing water vapor. In urban areas with minimal greenery, natural cooling mechanism is weakened and pavement dominates.
El Paso faces a double burden where the UHI effect amplifies the naturally high temperatures of the Chihuahuan Desert.
- Nighttime temperature spikes: The surrounding desert may drop in temperature after sunset, but urban pockets in the city stay significantly warmer. That loss of nighttime cooling can prevent the human body from recovering from daytime heat stress.
- Geographic trapping: El Paso's topography, such as the Franklin Mountains and the surrounding basin, can trap warm, stagnant air over developed areas, creating a heat-dome effect that lingers over densely populated neighborhoods.
- Increased ozone formation: Heat acts as a catalyst for chemical reactions between vehicle exhaust and industrial emissions. Higher temperatures from the UHI effect can contribute to ground-level ozone (smog), a major respiratory irritant.
- Disproportionate neighborhood impact: Research into El Paso heat equity shows that lower-income neighborhoods often have less tree canopy and more heat-absorbing surfaces. Those areas can be 5 to 10°F hotter than greener, wealthier parts of the city.
- Energy grid strain: As the heat island effect raises local temperatures, demand for air conditioning rises. That can create a feedback loop: higher demand leads to more power plant emissions and more waste heat from AC condensers, further warming the city.
A March 2026 study by the University of Cambridge identified a 'Data Heat Island' effect. AI data centers increase surrounding land surface temperatures by an average of 3.6°F. In extreme conditions, particularly in arid climates, surface temperatures can spike by more than 16°F
- Average signal: Mean LST increase near AI data centers after start of operations is about 3.6°F; the paper estimates more than 340 million people could be affected by the associated warming globally.
- Arid extremes: The largest uplifts appear in dry climates; the work reports extreme cases above 16°F for land surface temperature, with peak single-site LST change up to about 16.4°F (9.1°C).
- Beyond the property line: Elevated LST is not confined to the campus; the analysis finds a footprint on the order of 6.2 miles (10 km) from facilities.
- Surfaces and nighttime retention: Large LST changes reflect materials such as roofs and pavement that store solar energy by day and release it at night, which can disrupt the overnight cooling cycle that matters for public health.
Data centers and El Paso's heat: Large-scale data centers add a new variable to the local UHI. These facilities use massive industrial cooling to dissipate heat from thousands of servers. In a desert climate, that can add a constant stream of hot, dry air near the site and potentially extend the duration and intensity of heat-island conditions for nearby residents.
The Marinoni et al. results suggest data heat island signals can be especially large in arid settings and extend many miles beyond the fence. This adds cumulative stress on top of regional UHI and desert heat, not a neutral background. Permits and community benefit agreements should plan for LST and heat index near homes, not only property-line rules.
Cooling equipment, generators, and neighbor distance
AHUs, towers
SourcesWhere measured
EnforcementInversions
WeatherAir handling units and cooling towers generate constant low-frequency noise. This sound travels further and penetrates walls more easily than high-frequency noise. Residents in other desert markets describe the sound as a constant drone.
The City of El Paso states on its Data Center Resource Hub that data centers must comply with Title 9.40 of the El Paso City Code. In industrial zones the maximum sound level is 70 to 75 decibels (dBA) at the property line. For facilities near residential areas the limit is lower: 50 dBA at night. When homes are at stake, the decibel limit that protects you is checked at the receiving property line: the resident's lot line, where sound is measured outside your home, not at the operator's fence. Title 9.40 defines when the test applies at the neighbor's line versus the industrial line; read the chapter for the exact wording.
Those numbers describe what the code can permit. They should not be read as a guarantee of peace next door. A facility can stay within industrial fence-line limits while still producing a round-the-clock hum that carries into yards and bedrooms. Public health guidance often points to lower night noise for undisturbed sleep than a single headline figure suggests. Neighbors deserve clarity that compliance with Title 9.40 of the El Paso City Code is a legal floor, not proof that a hyperscale site will feel acceptable at the property line.
Arizona has updated municipal standards in the Phoenix area toward lower decibel limits and measurement at the data center property line (the source fence), so the plant must meet stricter limits where it sits. That contrasts with rules that only chase complaints at a distant lot without tightening the facility itself. Thermal inversions common in desert winters can bend sound downward, carrying it into distant neighborhoods.
Because data center noise is constant and often contains low-frequency tones, it can be more disruptive than intermittent city noises like traffic.
- Sleep disturbances: The persistent hum can lead to chronic insomnia, fragmented sleep, and nightmares. Unlike white noise, the mechanical nature of the sound can prevent the brain from entering deep sleep cycles.
- Cardiovascular issues: Long-term exposure to noise above 65 decibels is linked to increased blood pressure (hypertension) and elevated heart rates, which are risk factors for heart disease.
- Neurological and cognitive symptoms: Residents near facilities in Texas and Arkansas have reported frequent migraines, vertigo (dizziness), and nausea. The low-frequency pressure waves can affect the inner ear and vestibular system.
- Mental health: Chronic noise pollution is a significant stressor that can lead to increased anxiety, irritability, and a general decrease in quality of life.
- Hearing-related issues: Constant exposure can cause tinnitus-like symptoms, where residents perceive a ringing or buzzing in the ears even when away from the site.
Wildlife relies heavily on the "acoustical environment" for survival. A constant 24/7 hum creates a masking effect that drowns out natural sounds.
- Communication interference: Many species (birds, frogs, and insects) use specific frequencies to attract mates. The data center hum can mask these calls, leading to reduced reproductive success.
- Predator-prey dynamics: Predators (like owls or bats) that rely on hearing to locate prey find it harder to hunt. Conversely, prey animals may not hear a predator approaching over the mechanical noise.
- Altered behavior: Some birds have been observed changing the pitch of their songs or switching to nocturnal singing to be heard over industrial noise, which drains their energy reserves.
- Physiological stress: Animals exposed to chronic industrial noise show elevated levels of stress hormones (corticosterone), which can weaken their immune systems and reduce their overall lifespan.
- Habitat abandonment: Sensitive species may completely abandon otherwise suitable habitats to avoid the noise, leading to a loss of local biodiversity.
Independent monitoring and objective caps at the line named in the permit beat vague comfort standards.
| Jurisdiction | Day | Night | Where measured | Notes |
|---|---|---|---|---|
Arizona / Phoenix-area direction | 55 dBA | 45 dBA | Data Center Property line | Updated after residents complained about Data Centernoise |
El Paso (Title 9.40 of the El Paso City Code) | 70–75 dBA (industrial zone) | 50 dBA (near residential) | Reesident property line (homes) | Hub summarizes property-line tests; limits that protect neighbors are enforced at the receiving (resident) lot line where Title 9.40 applies |
Proposed Advocacy benchmark (pi.que llc) | 50 dBA | 45 dBA | Should be at the data center property line | Proposed standard by pi.que llc |
Capital intensity, hiring, and clawbacks
~$10B
Investment (reported)50
Jobs guaranteed* in contract~$110M
Incentives (order of magnitude)Reports indicate a $10 billion investment for the El Paso site, with company statements in the press floating on the order of 300 new jobs. That figure is a public promise, not a contractually guaranteed employment level. Research suggests permanent operational roles often total less than 20% of peak construction employment. Data centers are capital-intensive rather than labor-intensive.
El Paso has offered roughly $110 million in incentives. If that promise is taken at face value, the order of magnitude is hundreds of thousands of dollars per promised job; the number rises if realized permanent jobs end up far below the headline. Clawback clauses are legal tools that recover public funds if job or environmental targets are missed. Independent watchdogs recommend tying tax abatements to verifiable outcomes.
| Project | Investment (reported) | Promised jobs | Actual / revised | Notes |
|---|---|---|---|---|
Meta — Louisiana campus | $10B | 300–500 | ~100 | Local business press |
Stream Data Centers — New York | $11.2B | 125 | - | Jobs will cost $11 million each in lost tax revenue, one of the highest public costs Good Jobs First has ever seen |
Meta — El Paso | $10B | ~300 (company promise, only 50 in contract) | — | Actual permanent jobs TBD |
The 380 Meta Agreement: Decoding the fine print
50 Jobs
Job Floor$7.5M+
Road Subsidy35 Years
Lock-inThe 380 Economic Development Program Agreement is the primary contract between the City of El Paso and Wurldwide LLC. While public narratives focus on growth, the document establishes a 35-year framework of subsidies and restricted local authority. The City represents that the agreement is a valid and binding obligation.
One major discrepancy involves employment figures. Press releases often suggest hundreds of roles, yet the contract only requires a minimum of 50 full-time jobs. This requirement does not take effect until four years after the project reaches completion. If the job count slips, the City's only remedy is a reduction in grant amounts.
Financial transparency is another concern regarding public infrastructure. Meta must build a portion of Stan Roberts Sr. Avenue, but the City is obligated to reimburse Meta up to $7,500,000. This cap can increase by 20% if costs exceed initial estimates. Once finished, the City assumes all future maintenance costs for the road.
The contract includes several clauses that limit the City's power to govern the site or adapt to economic changes.
- The Tax Muzzle: Section 14.4 prohibits City staff from recommending or supporting any new taxes that specifically target the data center or the industry.
- The Regulatory Bubble: Section 14.2 prevents the City from modifying zoning or building regulations in a way that adversely affects the project.
- The 'Zombie' Extension: Section 5.2 states that for every year the City chooses not to appropriate funds for a grant, the contract term automatically extends by one year.
- No Public Bidding: Section 4.6.4 waives competitive bidding requirements for the road project, allowing Meta to select its own contractors using public funds.
The City's ability to recover value or penalize failures is strictly limited by the contract language.
- Sole Remedy: For most breaches, the City's sole and exclusive remedy is to terminate the agreement.
- No Refunds: Section 6.11 prevents the City from clawing back grants already paid, except in cases of specific immigration law violations.
- No Cross-Defaults: Section 6.12 ensures that a failure in one phase of the project cannot be used to cancel or penalize other successful phases.
- Fee Waivers: Section 5.3 waives nearly all impact, permit, and development fees, removing an immediate source of revenue for city services.
Transparency and the 'Black Box' effect
1 Per Year
Inspections30 Days
NoticeSection 19
Legal ShieldPublic oversight of a hyperscale facility is restricted by the 380 agreement. Section 4.7 limits the City to one inspection per year. The company requires 30 days of written notice before any official can enter the property. These visits must happen during normal business hours. They also require a security escort provided by the operator. This creates a controlled environment for any audit of water or energy use.
The contract also imposes secrecy on City employees via Section 4.7. Any individual inspecting the site can be forced to sign a confidentiality agreement. The company has the right to demand that any notes or documents viewed by City representatives stay at the project site. This prevents the creation of a public paper trail regarding internal operations or equipment performance.
Section 19 governs how public information requests are handled. The City is contractually obligated to oppose the release of any information the company labels as confidential. This includes submitting legal briefs to the Texas Attorney General to prevent disclosure. The company pays the legal costs. However, the City acts as the legal barrier to public record access.
When a private company controls the flow of information, the public cannot verify compliance with environmental or job targets.
- Restricted Access: The one visit per year rule in Section 4.7 prevents the City from conducting surprise audits of water meters or noise levels.
- Escorted Tours: Security escorts and pre-approved paths under Section 4.7 ensure that City staff only see what the company chooses to show.
- Legal Hurdles: Taxpayers seeking information through the Public Information Act face a City government that is legally bound by Section 19 to fight against them.
What El Paso can require
< 0.3 L/kWh
Water Intensity45 dBA (Night)
Noise LimitTiered System
AbatementsWater: Codify a city-wide prohibition on evaporative cooling towers. Prohibit the use of potable aquifer water for primary cooling functions to decouple growth from the drinking supply.
Energy: Reject proposals for dedicated fossil fuel plants like the McCloud facility. Require 24/7 hourly renewable matching utilizing local storage assets instead of annual credit accounting.
Air and Noise: Adopt a specific Data Center Noise Ordinance requiring acoustic louvers and encapsulated generators. Prohibit participation in fossil-fuel-based demand response programs for profit.
Governance: Implement a tiered incentive structure where full tax abatement requires meeting both job and environmental benchmarks. Publicly publish audit results on a municipal dashboard.
- Require facilities to be 'Liquid Ready' for future high-efficiency cooling.
- Mandate independent acoustic modeling and post-construction monitoring.
- Establish a citizen oversight board to review annual compliance.
- Require funding for a 'Virtual Power Plant' via residential solar and batteries.
Back to basics
If terms like PUE, hyperscale, or RECs are new, the Learn page walks through definitions with citations.
Go to Learn — basics and regional snapshotSpeak in public process
When utilities or councils post agendas, comments carry more weight when they cite docket numbers and measurable limits.
View city meetings