Who paysWhere capital can goWho pays nowData centresSite scoreboardThe hyperscaler overlapThe coalition
STEPS VENTURES · INDEPENDENT · SITE SCOREBOARD

Every site, costed and time-lined

31 data-center sites across 7 states drawing on Colorado River water, run by 24 operators. Open any site for its costed projects, capex, timeline and a viability rating grounded in that location's real power and water availability. The badge is how it cools today: water-free already there, evaporative the most to gain. The argument, and what any operator can do about it →

RegionAll (31)Arizona (CAP) (8)Nevada (SNWA) (3)Utah (5)Wyoming (1)Colorado Front Range (6)New Mexico (San Juan-Chama) (2)California Inland Empire (6)
CoolingAll (31)Water-free / closed-loop (14)Evaporative / water-cooled (5)Cooling not reported (12)
Find

Arizona (CAP) 8 sites

Google Mesa, AZ Water-free / closed-loop

Cooling as reported: Air-cooled

Water: City of Mesa. The City of Mesa's water supply comes from three primary sources, including Colorado River water delivered through the Central Arizona Project · Colorado River sourced · Minimal, as the facility is air-cooled for its technical operations. The city of Mesa had previously guaranteed the delivery of nearly 1 million gallons per day for an earlier, water-cooled design

Water use (estimated): ~23–232 AF/yr (from ~138 MW × air-cooling water intensity; actual varies with efficiency & reuse)

Land for solar-thermal cooling: ~276 ac needed · 2,877 ac open (non-built) within 3 km, 2,687 ac already built — ample

Sewer authority: City of Mesa Water Resources Department, Industrial Pretreatment Section ref · blowdown records held here, obtainable by public-records request

Costed projects (capex · timeline · viability by local power & water):

ProjectCapexTimelineAF/yrViabilityPower / water at this location
On-Site Blackwater to Potable Reuse$1,000-2,000/AF2-7 yearsDependent on facility water use, likely <100MediumThe project's moderate energy needs can be met by the local grid or on-site solar, which is abundant in Mesa. [18, 7]
Arizona's regulations for Direct Potable Reuse (DPR), effective March 2025, allow for this type of project. [3, 5]
Volumetric Agricultural Fallowing Agreement$200-400/AF annually12-24 months to establish agreementScalable based on agreement sizeHighThis project requires no significant power as it is a transactional water rights agreement.
Fallowing agreements are an established practice in the region, with counterparties available in nearby irrigation districts using Colorado River water. [11, 47]
Municipal Water Reuse Infrastructure Investment$150M for 8,000 AF/yr (City of Mesa project)3-5 years8,000-10,000HighThe City of Mesa's existing and planned infrastructure is supported by its own electric utility and power agreements. [13, 23]
Mesa is actively expanding its reuse capacity, exchanging treated effluent with the Gila River Indian Community for potable water supplies. [14, 15, 24]
Gila River Watershed RestorationVaries greatly, ~$1.5M total project cost for recent grants3-10+ yearsNot directly quantifiable as water volumeLowThis project has minimal power requirements, mainly for equipment during restoration activities.
Collaborative restoration efforts are active on the Lower Gila River, but water yield benefits are indirect and not easily quantified for offset purposes. [33, 34]
Brackish Groundwater Desalination for Aquifer Recharge$1,500-2,000/AF5-10 years10,000+MediumThe high energy demand (1-2.5 kWh/m³) would require significant new power generation or procurement, straining a grid that is already tightening. [9, 23, 2]
Brackish groundwater is available in several Arizona planning areas, including near Mesa, but concentrate disposal is a major challenge. [8, 10, 37]
Meta Mesa, AZ Cooling not reported

Cooling as reported: Uses a cooling technology that is stated to be significantly more water-efficient than the industry standard and reuses water multiple times

Water: City of Mesa. The City of Mesa's water supply includes Colorado River water from the Central Arizona Project · Colorado River sourced · Not specified, but the company has invested in water restoration projects

Water use (estimated): ~756–4,537 AF/yr (from ~450 MW × unknown-cooling water intensity; actual varies with efficiency & reuse)

Land for solar-thermal cooling: ~900 ac needed · 4,202 ac open (non-built) within 3 km, 2,385 ac already built — ample

Sewer authority: City of Mesa Water Resources Department / Mesa wastewater system (municipal sewer utility serving Mesa, AZ; exact outfall/collection connection for this site is not publicly confirmed in the sources found) ref · blowdown records held here, obtainable by public-records request

Costed projects (capex · timeline · viability by local power & water):

ProjectCapexTimelineAF/yrViabilityPower / water at this location
On-Site Advanced Water Recycling$15-35M18-30 months30-60HighMesa is inside the broader Phoenix metro grid, where data centers can usually secure power but interconnection/timing can be the gating item; on-site recycling adds moderate load that is typically manageable given Arizona's strong solar resource and available utility-scale/onsite solar options.
This is the most location-fit option because Mesa/WRP effluent is a real source and Arizona allows nonpotable reuse under ADEQ rules; the City of Mesa and nearby municipalities already operate wastewater systems that can supply reclaimed water, making true on-site recycling and reuse technically and administratively feasible.
Agricultural Efficiency Partnership for Volumetric Credits$1-5M12-24 months20-50LowPower is not the main constraint here; the challenge is that this is mostly a transaction/verification program with limited incremental electrical demand and Phoenix-area grid access is generally workable.
This is weakly grounded in this specific location because a durable, bankable volumetric offset market for Colorado River agricultural savings is not an established Arizona municipal mechanism; CAP shortage rules and conservation programs exist, but counterparty, measurement, and credit permanence are the hard part, so offsets are uncertain and not as straightforward as reuse or DPR.
Public-Private Partnership for Municipal Direct Potable Reuse (DPR)$25-80M3-6 years40-90MediumThe Phoenix metro grid can support DPR treatment loads, but the project would need interconnection and water-treatment power redundancy; Arizona solar is strong, but DPR still requires firm grid-backed power for continuous operation.
Arizona has a real pathway for potable reuse, but this is a public-infrastructure project, not a simple private one: ADEQ has potable reuse regulations/frameworks, and Mesa has access to municipal wastewater, yet a new DPR partnership would require city utility participation, advanced treatment approval, and a long permitting/engineering path.
Conversion to 100% Air-Based Cooling$20-60M18-48 months60-120MediumAir cooling shifts the burden from water to electricity, which Mesa-area grid and Arizona solar can generally support, but peak summer cooling reliability, transformer capacity, and any extra IT-side power use must be checked carefully.
This is physically feasible anywhere, but at Mesa's hot climate it is operationally challenging for a hyperscale data center; it can eliminate most evaporative demand, yet may require significant redesign, more compressor/fan energy, and possibly reduced economizer hours, so water savings are real but the performance tradeoff is large.
Brackish Groundwater Desalination and Aquifer Recharge$30-90M3-5 years25-70LowEnergy demand is high for brackish desalination and recharge pumping, but the Phoenix-area grid can usually serve it if a suitable interconnection and backup supply are secured; solar can help offset cost but not the continuous treatment load.
There is some brackish groundwater in Arizona, but siting, water rights, disposal of concentrate, and recharge permitting make this much less straightforward than reclaimed water in Mesa; it is technically possible in the region, yet not a natural fit for a data center-led offset because the brackish source and recharge project would need a very specific local hydrogeology and permits.
Microsoft Goodyear & El Mirage, AZ Water-free / closed-loop

Cooling as reported: Adiabatic cooling, which uses outside air instead of water for cooling when temperatures are below 85 degrees Fahrenheit. For future data centers in Goodyear, Microsoft has agreed to switch to air-cooled systems

Water: City of Goodyear. Goodyear's water portfolio includes both groundwater and surface water from the Colorado River delivered by the Central Arizona Project · Colorado River sourced · An analysis estimated the Goodyear data center's water use at around 56 million

Water use (estimated): ~24–240 AF/yr (from ~143 MW × air-cooling water intensity; actual varies with efficiency & reuse)

Land for solar-thermal cooling: ~286 ac needed · 3,232 ac open (non-built) within 3 km, 1,766 ac already built — ample

Sewer authority: City of Goodyear · sewer discharge: 1.2 MGD capacity at full build-out (City of Goodyear 157th Ave WRF, AZPDES AZ0022357); Microsoft since agreed to air-cooled source · blowdown records held here, obtainable by public-records request

Costed projects (capex · timeline · viability by local power & water):

ProjectCapexTimelineAF/yrViabilityPower / water at this location
Cooling System Retrofits$1.7M per 1 MW of cooling capacity6-18 months56-179 (estimated based on 56M gallon reported use)HighThe local grid, supported by abundant solar resources and new battery storage, can handle the increased power load from air-cooled systems. [8, 24]
This project eliminates the need for a water source for cooling, aligning with Microsoft's agreement with the City of Goodyear. [30]
Local Direct Potable Reuse (DPR)$1B for a 60 MGD facility (Phoenix example)5-7 years (design, permitting, construction)Varies with investment; a 60 MGD plant yields ~67,000 AF/yrMediumDPR is energy-intensive, requiring significant and reliable power from a grid already supporting numerous data centers and a growing population. [33, 8]
Goodyear is expanding its wastewater reclamation facilities, and Arizona's new AWP rules (effective March 2025) provide a clear regulatory path for DPR projects. [18, 3, 9, 14]
Agricultural Fallowing Credits$150-400/AF (based on various programs)12-24 months to establish agreementsScalable based on number of agreementsHighThis project has no direct impact on local power infrastructure as it is a financial and water management transaction.
Active fallowing programs exist within the Central Arizona Project and with parties like the Gila River Indian Community, making counterparties available.
On-Site Blackwater/Graywater Reuse$1.2M for a 4,000 GPD system (SFPUC example)18-36 months1-5 (dependent on building scale and occupancy)LowThe power demand is minimal and easily supported by existing site infrastructure and the local grid. [29]
While technically feasible, the small water yield makes it a low-impact solution for a data center's large-scale water-use mitigation goals. [38, 44]
Watershed Restoration~$2,500/acre for thinning (yields ~0.07 AF/acre)Ongoing, with benefits realized over many years~0.07 AF per acre treated in the first yearMediumThis project has no direct impact on the Goodyear power grid.
Partnerships with entities like SRP are available, but the water yield per acre is low, transient, and geographically dispersed, making it difficult to quantify for a specific offset. [4, 6, 7]
CyrusOne Chandler, AZ Cooling not reported

Cooling as reported: Zero-water cooling

Water: City of Chandler. The city's water supply is a mix from the Salt and Verde rivers, and the Colorado River · Colorado River sourced · Negligible for cooling, with only about 180,000 gallons a year for humidification · solar-thermal: ~338 ac needed

Water use (estimated): ~1,420–2,556 AF/yr (from ~169 MW × evap-cooling water intensity; actual varies with efficiency & reuse)

Land for solar-thermal cooling: ~338 ac needed · 2,355 ac open (non-built) within 3 km, 3,582 ac already built — ample

EPA facility record: CYRUSONE - CHANDLER (Air permit AZMCA004013X132941) · ECHO report

Sewer authority: City of Chandler Water Utilities Department / City of Chandler wastewater system (served by the City’s public sewer collection and treatment system; no public record found in the sources reviewed tying this specific parcel to a regional sanitation district) ref · blowdown records held here, obtainable by public-records request

Costed projects (capex · timeline · viability by local power & water):

ProjectCapexTimelineAF/yrViabilityPower / water at this location
On-Site Water Capture for Operational Needs$0.5-2M6-18 months0.1-0.2LowPower is not the constraint; Chandler sits in a strong Phoenix-area grid with ample utility service for this load, and onsite capture systems are small electrical loads. Solar resource is strong, but it does not materially change the economics here.
This site already uses zero-water cooling and only about 180,000 gallons/yr for humidification, so there is very little water to capture or offset; Arizona reuse rules can support nonpotable reuse, but the volume available at this facility is too small to matter for net Colorado River reduction.
VWBA for Agricultural Fallowing$150-400/AF of annualized water benefit6-18 months500-5,000+HighGrid capacity is not a limiting factor for a paper water-benefit contract; Chandler's desert-southwest grid is mature and this project is not power-intensive.
Agricultural fallowing offsets are regionally plausible in Arizona because Colorado River water is heavily used in Central Arizona agriculture and the state has established water-rights/accounting pathways under ADWR/Colorado River management, though each deal needs a real counterparty and verification that the conserved water is actually reduced consumptive use.
Investment in Local Watershed Restoration$200-800/AF of estimated benefit12-36 months50-500MediumPower availability is irrelevant to watershed restoration; Chandler's grid and strong solar resource do not affect feasibility.
Restoration can improve local hydrology and resilience, but in Arizona it usually does not create a clean, durable, legally countable Colorado River consumptive-use offset unless paired with a defined accounting framework; reuse/DPR rules do not substitute for a verified river-offset counterparty.
Financing Local Turf-to-Xeriscape Conversions$100-300/AF of estimated annual savings6-24 months100-1,000MediumThis is not power constrained; the local grid can easily support administration/monitoring, and high solar resource is largely irrelevant to the water accounting.
Arizona municipalities, including Chandler-area users, can implement landscape conversion programs, but the water saved is usually local potable/municipal demand reduction rather than a direct Colorado River entitlement transfer, so the project needs careful accounting to qualify as a net Colorado River benefit.
Capital Contribution to Municipal Aquifer Recharge$300-1,200/AF of credited recharge18-48 months200-2,000HighGrid capacity is ample for a financing/participation role, and the local solar resource is immaterial to the creditability of recharge.
This is one of the most grounded options in the Phoenix metro because Arizona has active aquifer recharge/assured supply infrastructure and reclaimed-water systems, but the project must use approved recharge facilities and credit rules; it is most credible when tied to actual Arizona water banking or municipal reuse volumes rather than vague 'restoration' claims.
Solar-Thermal Chiller$540-1,350M for ~338 acres at $2-4M/acre installed36-72 monthsNegligible direct water; roughly 0.2-0.3 if replacing cooling tower makeup from humidification-equivalent needsLowChandler has very strong solar irradiance and the Phoenix metro grid can interconnect large loads, but a ~169 MW data center needing ~338 acres of collectors is land-hungry and operationally mismatched to a data center that already has zero-water cooling.
Water is not the problem at this site: the data center already reports negligible cooling water use, and Arizona solar-thermal cooling would not create a meaningful additional Colorado River reduction; moreover, the land requirement is difficult near Chandler, where built-up parcels dominate and nearby desert parcels are limited.
Vantage Data Centers Goodyear, AZ Cooling not reported

Cooling as reported: Not specified, but the company emphasizes sustainable design

Water: City of Goodyear. Goodyear's water supply includes Colorado River water · Colorado River sourced · Not publicly reported · solar-thermal: ~352 ac needed

Water use (estimated): ~296–1,774 AF/yr (from ~176 MW × unknown-cooling water intensity; actual varies with efficiency & reuse)

Land for solar-thermal cooling: ~352 ac needed · 2,831 ac open (non-built) within 3 km, 3,081 ac already built — ample

Sewer authority: City of Goodyear ref · blowdown records held here, obtainable by public-records request

Costed projects (capex · timeline · viability by local power & water):

ProjectCapexTimelineAF/yrViabilityPower / water at this location
Fund Agricultural Efficiency with Volumetric Water Credits$630-$750/AF6-12 monthsScalable with investmentHighThis is a financial transaction and has no direct impact on local power infrastructure.
Partnerships are feasible as state-funded programs to improve on-farm irrigation efficiency are active and oversubscribed in Arizona. [18, 22, 23]
Invest in a City of Goodyear Recycled Water ProjectProject-dependent, likely multi-million dollar contribution3-5 yearsDependent on project scale and investment levelHighThe project's power needs are part of the city's infrastructure planning and not a direct burden on the data center's interconnection.
Goodyear has a robust recycled water program, produces A+ effluent, and is actively expanding its reclamation facilities and reuse capabilities. [2, 5, 14, 31, 35]
Purchase and Retire Agricultural Water RightsHighly variable; potentially $6,000-$12,000+/AF for senior rights12-24 monthsDependent on volume of rights purchasedMediumThis is a legal/financial transaction with no direct impact on local power infrastructure.
While transferable rights exist, the process is complex, requires ADWR approval, and may face opposition, making willing sellers and timely execution uncertain. [19, 26, 41]
Fund Off-site Aquifer Recharge with Non-CR WaterProject-dependent, likely multi-million dollar contribution2-4 yearsDependent on project scale and water availabilityMediumPumping for recharge requires power, but it is a manageable load within the local grid's capabilities.
Goodyear actively recharges the aquifer with surplus CAP water and effluent, indicating geological suitability and an existing operational framework. [2, 5, 14]
Develop On-site Blackwater Treatment and Reuse$3-7M for a small-scale commercial MBR system24-36 months<15 AF/yr (estimated)LowThe system's power demand is negligible relative to the data center's total IT load.
While Arizona's regulations are becoming more favorable for reuse, the volume of blackwater from staff is insignificant compared to the cooling demand of a 176MW data center. [10, 11, 33]
SOLAR-THERMAL CHILLER$700M - $1.4B (at $2-4M/acre for 352 acres)4-6 yearsHigh avoidance, but displaces existing low-water cooling systemLowThe excellent solar resource in Goodyear is ideal, but the project primarily offsets thermal load, not the main grid interconnection capacity challenge. [9, 34, 46]
The massive land requirement (352 acres) is a major obstacle, despite claims that land is 'likely' available and Vantage's current design already uses virtually no water for cooling. [6, 17, 21]
Digital Realty Chandler, AZ Evaporative / water-cooled

Cooling as reported: The company prioritizes water efficiency, including the use of dry cooling systems that avoid water evaporation

Water: City of Chandler. The city's water supply includes about 37% from the Colorado River · Colorado River sourced · Estimated at 0.7 million gallons per day at full load · solar-thermal: ~108 ac needed

Water use (estimated): ~454–817 AF/yr (from ~54 MW × evap-cooling water intensity; actual varies with efficiency & reuse)

Land for solar-thermal cooling: ~108 ac needed · 852 ac open (non-built) within 3 km, 5,313 ac already built — ample

Sewer authority: City of Chandler – Public Works & Utilities (municipal wastewater collection/treatment; regional disposal via the City of Chandler POTW/WRF system, not a separate regional sanitation district) ref · blowdown records held here, obtainable by public-records request

Costed projects (capex · timeline · viability by local power & water):

ProjectCapexTimelineAF/yrViabilityPower / water at this location
Fund Agricultural Fallowing for Water Benefit Credits$2-8M upfront for a meaningful pilot/credit package; roughly $200-800/AF depending on contract term and transaction structure12-36 months50-300MediumNo material power requirement beyond transaction/monitoring, so Chandler’s grid capacity is not a constraint; this is an offset project, not a load-serving project.
Arizona has active agricultural fallowing/forbearance mechanisms tied to Colorado River conservation, but credits are policy- and counterparty-dependent and do not directly create local municipal supply; the 37% Colorado River exposure makes this a plausible offset, not a physical replacement.
Retrofit to Use 100% Municipal Recycled Water$10-35M, or about $1,500-5,000/AF-year of firm service capacity depending on piping, storage, and treatment upgrades18-48 months228MediumChandler and the East Valley have strong grid access and the plant load is modest relative to metro infrastructure, so electrical interconnection is feasible; the main power issue is continuous pumping/treatment, not supply.
Recycled effluent is actually available in this metro via municipal reclaimed-water systems, but 100% dedicated delivery to a data center depends on line proximity, purple-pipe capacity, utility permitting, and water-quality specs; this is more feasible than DPR here, but not trivial.
Invest in a Municipal Direct Potable Reuse (DPR) Project$20-60M contribution/share of a larger utility DPR program; private cost is highly program-specific and typically much higher per AF than nonpotable reuse36-72 months228LowThe grid can support advanced treatment loads, but DPR depends on utility-scale infrastructure and long permitting timelines rather than local data-center power availability.
Arizona allows reuse under a regulated framework, but potable reuse in the Phoenix-metro area is still utility-led and not a near-term bespoke option for a single industrial customer; there is no ready-made DPR supply at the Chandler site today.
On-Site Blackwater Treatment and Aquifer Recharge$15-40M for treatment, storage, injection/recharge infrastructure, permitting, and monitoring24-60 months228LowThe site’s electrical service can likely support an onsite treatment plant, but that is not the limiting factor; the bottleneck is regulatory approval and hydrogeologic feasibility for recharge.
Arizona recharge and reclaimed-water rules are real, but a data center blackwater-to-recharge scheme faces major permitting, siting, and groundwater-basin requirements; in a built-up Chandler context, this is unlikely to be the fastest or cleanest path to zero Colorado River dependence.
Fund Upstream Watershed Restoration in the Salt-Verde System$1-10M initial funding, or roughly $50-300/AF-equivalent depending on watershed model and crediting assumptions12-48 months20-200LowPower availability is irrelevant; this is a nature-based offset, not a direct utility supply project.
Salt-Verde watershed restoration can improve system resilience, but it does not generate a fungible, legally recognized Colorado River water entitlement for Chandler; crediting against a specific data-center withdrawal is uncertain and usually indirect.
Solar-Thermal Chiller$216-432M for ~108 acres of collectors at roughly $2-4M/acre installed, excluding major storage, land assembly, and integration costs24-60 months228LowArizona has excellent solar resource, but a solar-thermal cooling field sized for a ~54 MW site is land-intensive and adds major thermal/electrical integration complexity versus simply buying grid power or using dry cooling; the local grid can supply the load more directly.
This does not solve water sourcing by itself and the required 108 acres is hard to assemble near a built-up Chandler data center; while some nearby desert parcels exist and solar irradiance is strong, the land footprint and capital intensity make it a poor fit for this specific site.
QTS Data Centers Phoenix & Glendale, AZ Cooling not reported

Cooling as reported: Not specified

Water: City of Phoenix / City of Glendale. The City of Phoenix's water supplies include the Colorado River · Colorado River sourced · Not publicly reported

Water use (estimated): ~1,462–8,771 AF/yr (from ~870 MW × unknown-cooling water intensity; actual varies with efficiency & reuse)

Land for solar-thermal cooling: ~1,740 ac needed · 3,769 ac open (non-built) within 3 km, 2,425 ac already built — feasible

Sewer authority: City of Glendale Water Services Department ref · blowdown records held here, obtainable by public-records request

Costed projects (capex · timeline · viability by local power & water):

ProjectCapexTimelineAF/yrViabilityPower / water at this location
Agricultural Water Offset$140-$420/AF6-18 months200HighThis project has no direct power requirement, avoiding any new load on the local grid.
Counterparties are available as agriculture uses over 70% of the region's Colorado River allocation, and established programs offer a framework for these agreements. [8, 14, 25, 47]
On-Site Recycled Water Conversion$1-3M24-36 months200HighThe power required for on-site pumping and potential polishing treatment is minor and can be supported by existing grid connections or on-site solar. [7, 9]
Both Phoenix and Glendale operate water reclamation facilities and encourage the use of recycled water for industrial purposes like cooling. [29, 33, 35, 39]
Air-Cooled System Retrofit$14-15M18-30 months200MediumThis significantly increases electricity demand, which the robust Phoenix-area grid and solar resources can support, but at a high operational cost. [22, 44, 45]
This project eliminates the need for an external water source for cooling, directly achieving the water-saving goal on-site. [37, 42]
Brackish Groundwater Desalination$4-8M36-60 months200LowThe reverse osmosis process is energy-intensive, requiring a significant new power load that would likely necessitate a dedicated substation and increase operational costs. [4, 16]
While brackish aquifers exist in the West Salt River Valley, permitting for wells and especially for the disposal of brine concentrate is complex, costly, and a major logistical hurdle in the region. [6, 19, 40, 51]
Stream Data Centers Goodyear, AZ Water-free / closed-loop

Cooling as reported: Air-cooled chillers with a closed-loop water system

Water: City of Goodyear. Goodyear's water supply includes Colorado River water · Colorado River sourced · The closed-loop system is filled once and does not require additional water, making its daily use comparable to an office building of similar size · solar-thermal: ~560 ac needed

Water use (estimated): ~7–68 AF/yr (from ~40.5 MW × air-cooling water intensity; actual varies with efficiency & reuse)

Land for solar-thermal cooling: ~560 ac needed · 3,014 ac open (non-built) within 3 km, 2,108 ac already built — ample

Sewer authority: City of Goodyear Water Services / wastewater utility (local municipal sewer service for Goodyear, AZ; exact POTW receiving this parcel’s flows is not publicly confirmed in the sources I could verify) · blowdown records held here, obtainable by public-records request

Costed projects (capex · timeline · viability by local power & water):

ProjectCapexTimelineAF/yrViabilityPower / water at this location
Fund Agricultural Fallowing (Water Benefit Agreement)$2-8M initial buy-in or ~$1,500-4,000/AF-year of conserved Colorado River water6-18 months500-2,000MediumPower is not the constraint here; this is a financial water-rights/offset transaction, and Goodyear sits in the fast-growing West Valley with ample grid access but not a project-dependent electrical need.
Arizona has active Colorado River conservation and Arizona Water Banking/tribal/irrigation programs, but a true one-for-one offset requires a willing counterparty with transferable conserved Colorado River consumptive use and durable accounting; that is feasible in principle in Arizona but not guaranteed locally and is more policy/contract dependent than physically water-supply dependent.
Invest in Local Watershed Restoration$1-5M12-36 months0-100LowNo meaningful power dependency, and Goodyear's grid/solar conditions do not materially improve a watershed-restoration offset because the project does not create firm water supply.
This is not a credible path to zero net Colorado River water for an Arizona data center because watershed restoration typically yields uncertain, non-firm, and often non-transferable hydrologic benefits; Arizona water law does not treat generic restoration as a usable replacement for Colorado River consumptive use.
Sponsor Municipal Recycled Water Expansion$10-40M contribution or $1,000-3,000/AF-year of firmed supply2-5 years300-1,500MediumGrid availability is adequate for the data center, but reclaimed-water expansion is a municipal civil-works project, not a power-limited one; the main constraint is pipeline/service-area access and treatment capacity.
Goodyear is in the Phoenix metro where reclaimed water exists and municipal reuse is common, but the utility service area, purple-pipe reach, and Arizona reclaimed-water rules mean a dedicated supply to this site would likely require new mains, agreements, and possibly indirect-use approvals; physically available in the region, but not instantly at this parcel.
Develop On-Site Blackwater to Potable Reuse$20-60M+3-7 years20-150LowThe local grid can power the treatment plant, but this project is constrained by treatment complexity and permitting, not by electricity supply; interconnection is likely far easier than the water/wastewater approvals.
Arizona permits potable reuse through ADWR and ADEQ pathways, but an on-site blackwater-to-potable system for a private data center is highly unusual, would face major public-health/permitting hurdles, and the site’s current air-cooled configuration produces too little wastewater to make meaningful firm supply at scale.
Fund Brackish Groundwater Recharge Project$8-25M2-5 years200-1,000LowPower is available, but brackish groundwater projects are more limited by hydrogeology, treatment/disposal, and recharge permitting than by electricity or solar resource.
There is brackish groundwater in parts of Arizona, but a recharge project only helps if there is a suitable aquifer, water-rights/accounting pathway, and an entity able to use the credits; at Goodyear, this is not a straightforward local offset because the site’s immediate service area is on the CAP/municipal supply side rather than a simple brackish-aquifer banking market.
Solar-Thermal Chiller$1.1B-2.2B for ~560 acres of collectors at $2-4M/acre; plus site/building integration4-8 yearsPotentially large on paper, but not a practical zero-water solution for this siteLowThe West Valley has excellent solar resource, but a 280 MW data center would need an enormous collector field and thermal storage to match 24/7 cooling demand; grid interconnection for the load is possible in metro Phoenix, yet the solar-thermal plant would be a massive land- and capital-intensive adjunct rather than a simple site-scale solution.
The immediate Goodyear/far-west Phoenix area does have strong solar insolation and available desert land in some corridors, but ~560 acres of contiguous collector field is a very large footprint for an exurban site, and solar-thermal cooling is not a normal, bankable way to offset Colorado River consumptive use in Arizona; it reduces electric load more than it directly solves the water accounting problem.

Nevada (SNWA) 3 sites

Google Henderson, NV Evaporative / water-cooled

Cooling as reported: Evaporative cooling. The Southern Nevada Water Authority has since banned this method for new data centers

Water: City of Henderson, supplied by the Southern Nevada Water Authority (SNWA) · Colorado River sourced · Estimated at 352 million gallons in 2024. Another report cited 205 million gallons for cooling in 2024

Water use (estimated): ~1,008–1,815 AF/yr (from ~120 MW × evap-cooling water intensity; actual varies with efficiency & reuse)

Reported: ~630 AF/yr government

Land for solar-thermal cooling: ~240 ac needed · 2,101 ac open (non-built) within 3 km, 4,565 ac already built — ample

Sewer authority: City of Henderson Department of Utility Services ref · blowdown records held here, obtainable by public-records request

Costed projects (capex · timeline · viability by local power & water):

ProjectCapexTimelineAF/yrViabilityPower / water at this location
Cooling System Retrofit$7M-$15M12-24 months~629MediumA retrofit to air or closed-loop liquid cooling would increase electricity demand on a grid already strained by data center growth, though NV Energy is actively planning major expansions. [33, 43]
This directly addresses the ~205 million gallons used for cooling and aligns with SNWA's ban on evaporative cooling for new data centers, making it a highly relevant solution. [33]
SNWA Potable Reuse ProgramContribution-based (e.g., SNWA contributing $750M to a regional project)3-5+ years for new large-scale facilities~1,080HighThis municipal-level project has no direct power impact on the data center, though water recycling and pumping at this scale is energy-intensive for the utility. [20]
SNWA is actively pursuing direct potable reuse, and Nevada has established regulations (Category A+) for it, making this a structurally sound and available long-term option. [29, 35]
Acquisition of VWBAs$150-$400/AF (market variable)Immediate (contract-dependent)~1,080HighThis is a financial transaction with no direct impact on the data center's power consumption.
VWBAs are an established market mechanism for offsetting water use, and organizations like the Bonneville Environmental Foundation facilitate such projects in the Colorado River basin. [21, 41, 48]
On-Site Blackwater/Greywater Recycling$1M-$3M for a commercial system18-24 months< 5LowThe system requires additional power, but the load would be minor compared to the data center's total consumption.
While Nevada has regulations for non-potable reuse, data centers generate minimal wastewater, making this approach insufficient to offset significant cooling or operational water use. [13, 19]
Funding Las Vegas Wash RestorationContribution-based (e.g., portions of multi-million dollar federal grants)Ongoing (project-dependent)~1,080HighThis is a financial contribution to an external environmental project and has no direct power impact on the data center.
The Las Vegas Wash is the primary mechanism for SNWA's return-flow credits, making investment here a direct and proven way to return water to Lake Mead. [4, 5, 6, 9]
Switch Las Vegas, NV Water-free / closed-loop

Cooling as reported: Primarily evaporative cooling, but the company also utilizes closed-loop systems and has patents on innovative cooling technologies suited for the desert climate

Water: Las Vegas Valley Water District (LVVWD), part of the SNWA · Colorado River sourced · Not specified, but Switch is listed as a large water user in the region

Water use (estimated): ~53–529 AF/yr (from ~315 MW × air-cooling water intensity; actual varies with efficiency & reuse)

Land for solar-thermal cooling: ~630 ac needed · 2,396 ac open (non-built) within 3 km, 4,428 ac already built — ample

EPA facility record: SWITCH NAP 11 (NPDES NVCS43114) · ECHO report

Sewer authority: Southern Nevada Water Authority (SNWA) / City of Las Vegas sanitary sewer service area, with treatment handled by the regional wastewater system operated through the Las Vegas Valley Water District / Clark County regional sewer arrangements; the exact receiving POTW for this specific parcel is not publicly confirmed in a single primary source · blowdown records held here, obtainable by public-records request

Costed projects (capex · timeline · viability by local power & water):

ProjectCapexTimelineAF/yrViabilityPower / water at this location
On-Site Blackwater Recycling and Aquifer Recharge$20-60M24-48 months200-800MediumLVVWD/Nevada has ample grid access in metro Las Vegas, and the desert solar resource is strong; the bigger issue is not power but the continuous electrical load and backup needed for advanced treatment, pumping, and monitoring.
Las Vegas can use reclaimed water and Managed Aquifer Recharge is legally and physically supported in Southern Nevada, but true on-site blackwater-to-recharge is constrained by Nevada water-quality permitting and the need to keep recharge water compliant with SNWA/LVVWD and NDEP requirements; this is feasible only as a tightly controlled pilot, not a large-scale offset.
Financing SNWA Return-Flow Credits via Municipal Effluent$1-5M6-18 months500-2,000HighPower is not the binding constraint; Las Vegas has strong utility-service availability and very good solar resource if any associated pumping or treatment is co-located with PV.
This is one of the few real local offset pathways because SNWA/LVVWD already operate a mature reclaimed-water and return-flow-credit framework, and municipal effluent/reuse water is actually available in the Las Vegas Valley under existing reuse rules.
Volumetric Water Benefit Agreement (VWBA) for Agricultural Fallowing$0.1-0.5M6-24 months1,000-10,000MediumThe local grid and solar resource are sufficient; this project is mostly a contracting/verification exercise rather than an energy-intensive one.
The challenge is counterparty water availability, not local supply: Nevada has limited in-state agricultural offset opportunity at scale, so a credible VWBA would likely need Colorado River basin fallowing or conserved consumptive use outside the Las Vegas Valley, with rigorous additionality and accounting.
Full Conversion to Air-Cooled Systems via SNWA's WET Program$50-150M36-72 months3,000-15,000LowLas Vegas can physically support the electrical load only if the site secures large interconnection capacity and possibly on-site/contracted solar plus storage; however, air cooling materially increases peak demand in an already summer-stressed desert grid context.
The Water Efficient Technologies (WET) program exists locally, but full conversion of a hyperscale campus from evaporative to air-cooled systems is usually operationally difficult in the Mojave heat and may shift, not eliminate, water demand by increasing power use; it is not a clean, guaranteed zero-net Colorado River water solution.
Investment in Las Vegas Wash Watershed Restoration$0.5-3M6-24 months50-300LowPower availability is essentially irrelevant here because this is a restoration finance project, not an infrastructure-heavy one.
The Las Vegas Wash is a real local watershed and restoration target, but habitat/watershed restoration is not generally an accepted direct volumetric Colorado River water offset under Nevada/SNWA accounting, so it is weak for claiming zero-net water status even if it has environmental value.
Flexential North Las Vegas, NV Water-free / closed-loop

Cooling as reported: Uses a glycol compound for cooling loops and also air cooling

Water: City of North Las Vegas, supplied by the SNWA · Colorado River sourced · Around 20 million gallons in 2024

Water use (estimated): ~2–15 AF/yr (from ~9 MW × air-cooling water intensity; actual varies with efficiency & reuse)

Land for solar-thermal cooling: ~18 ac needed · 3,619 ac open (non-built) within 3 km, 3,105 ac already built — ample

Sewer authority: City of North Las Vegas Utilities Department ref · blowdown records held here, obtainable by public-records request

Costed projects (capex · timeline · viability by local power & water):

ProjectCapexTimelineAF/yrViabilityPower / water at this location
Cooling System Upgrade to Waterless TechnologyUp to 50% of project cost covered by SNWA incentive. Total cost varies, but can be comparable to water-cooled systems. [38, 39]6-18 months for a facility retrofit. [14]~62HighAir-cooled systems can use more energy than evaporative cooling, but the local grid, heavily supplied by natural gas and a growing solar portfolio, can support the load. [19, 23, 45]
This project eliminates the need for process cooling water, making it highly viable in a water-scarce region. [33]
Investment in SNWA's Return-Flow Credit Programn/a (Financial transaction)Immediate upon agreement with SNWA.~62HighThere is no direct power impact on the data center as this is a financial transaction leveraging existing municipal water treatment infrastructure. [15]
The program is well-established; SNWA and its partners operate the wastewater treatment and return system that generates these credits. [15, 16]
Volumetric Water-Benefit Agreement (VWBA) for Agricultural Conservation$140-$417/AF. [1, 5, 8]6-24 months to identify partner, execute agreement, and verify savings.~62HighThis is an off-site financial agreement for conservation, imposing no additional power demand on the North Las Vegas grid.
Partnerships with agricultural water users in the Colorado River Basin are available, and this is considered a highly cost-effective method for conservation. [1, 5, 11]
On-Site Brackish Groundwater Desalination Plant$500 - $2,947 per m3/day of capacity (inflation-adjusted to 2020). [34] For ~62 AFY (~208 m3/day), this is roughly $0.1M - $0.6M.18-36 months for planning, permitting, and construction.~62LowDesalination is energy-intensive; while the local grid has capacity, the high consumption makes this a costly and power-demanding option. [6, 7, 23]
Requires sourcing a local brackish groundwater supply and, critically, a permitted and viable method for brine concentrate disposal, which is a major challenge.
Funding for Las Vegas Wash RestorationContribution-based (e.g., projects range from hundreds of thousands to millions of dollars). [2, 4]Immediate upon contribution.0 (direct); indirect benefitMediumThis is a financial contribution to an external environmental project and has no direct impact on the data center's power consumption.
This project improves the quality of water for which SNWA earns return-flow credits but does not generate new water credits for the funder. [16, 27, 37]

Utah 5 sites

Meta Eagle Mountain, UT Cooling not reported

Cooling as reported: Uses a cooling technology that is more water-efficient than the industry standard and reuses water multiple times

Water: Eagle Mountain City. The city collaborates with Meta on water reuse projects · Colorado River sourced · Over 35 million gallons in 2024

Water use (estimated): ~192–1,149 AF/yr (from ~114 MW × unknown-cooling water intensity; actual varies with efficiency & reuse)

Land for solar-thermal cooling: ~228 ac needed · 6,370 ac open (non-built) within 3 km, 49 ac already built — ample

Sewer authority: Eagle Mountain City wastewater/sewer utility (city-operated wastewater collection and treatment service; local public-works/utility function) · blowdown records held here, obtainable by public-records request

Costed projects (capex · timeline · viability by local power & water):

ProjectCapexTimelineAF/yrViabilityPower / water at this location
On-Site Direct Potable Reuse (DPR) Loop$8-20M24-48 months100-300MediumMeta’s Eagle Mountain campus can likely support a modular DPR plant electrically, and Utah has strong solar resource, but treatment/reliability gear still needs firm interconnection and backup power; the bigger issue is not electricity but regulatory/process complexity.
Utah has a direct potable reuse framework, but an on-site DPR loop inside a data center is still highly permitting- and public-health-review intensive; this site does have collaboration with Eagle Mountain City, which helps, but it is not a trivial path to meaningful Colorado River offset volume.
Agricultural Efficiency Water Lease via Utah Water Bank$0.5-2M6-18 months500-2,000LowThis option is essentially power-light and can be supported by the local grid easily; solar/power availability is not the constraint.
Utah has water banking and transfer mechanisms, but Eagle Mountain/Utah County is not an active, large-scale Colorado River agricultural offset market like some Basin programs, and finding a durable counterparty with legally transferable consumptive-use savings is the hard part.
Municipal Recycled Water Swap$3-10M18-36 months200-800MediumLocal electric service and Utah’s strong solar profile can support pumping and treatment reuse infrastructure without difficulty.
Eagle Mountain City has publicly indicated collaboration with Meta on water reuse, so a municipal reclaimed-water exchange is one of the more location-realistic options; however, the available volume depends on city wastewater flows and reuse infrastructure, which are still limited compared with large Colorado River offsets.
Brackish Groundwater Desalination$20-60M24-48 months300-1,500LowThe grid can serve a desal plant, but energy intensity is materially higher than reuse and the need for dependable pumping plus pretreatment favors a robust interconnection and backup supply.
Brackish groundwater is not a clearly established, ready local offset source in Eagle Mountain at the scale needed; even if a brackish aquifer were developed, the key problem is whether the produced water is truly additional and legally countable as Colorado River offset.
Upper Basin Watershed Restoration for Volumetric Benefits$1-8M12-60 months50-500LowPower availability is not the binding constraint; restoration projects are mostly labor/land/restoration funded rather than electricity constrained.
Watershed restoration can produce uncertain, hard-to-verify volumetric benefits, and in the Upper Basin those benefits are often not legally straightforward Colorado River consumptive-use offsets; this makes it a weak fit for a firm zero-net claim in Utah.
Novva Data Centers West Jordan, UT Water-free / closed-loop

Cooling as reported: Waterless cooling system

Water: West Jordan City · Colorado River sourced · 3 million gallons between October and September, with two-thirds used for landscaping. The company estimates a traditional evaporative cooling system for a campus of this size would use 300-400 million gallons annually · solar-thermal: ~400 ac needed

Water use (estimated): ~5–52 AF/yr (from ~31 MW × air-cooling water intensity; actual varies with efficiency & reuse)

Land for solar-thermal cooling: ~400 ac needed · 4,884 ac open (non-built) within 3 km, 1,634 ac already built — ample

Sewer authority: South Valley Water Reclamation Facility ref · blowdown records held here, obtainable by public-records request

Costed projects (capex · timeline · viability by local power & water):

ProjectCapexTimelineAF/yrViabilityPower / water at this location
On-Site Wastewater Treatment and Reuse$1-3M24-36 months3-9MediumThe power requirement is minor and can be easily supported by the local grid.
Utah DEQ regulations approve on-site reuse for non-potable uses like dust control and toilet flushing, making the 1M gallons of operational water (~3 AF) reusable. [14, 26]
Agricultural Fallowing for Volumetric Water Benefits$390/AF6-12 monthsVariableHighThis project has no power requirements.
Utah has an active pilot program for fallowing, and willing agricultural partners exist in the region, making offset agreements feasible. [7, 17]
Campus-Wide Xeriscaping and Smart Irrigation Conversion$8-16/sqft6-12 months4-6HighSmart irrigation controllers have minimal power needs easily met by existing infrastructure.
This project directly reduces the 2M gallons (~6 AF) used for landscaping; Utah offers financial incentives for these conversions. [12, 43]
Stormwater Capture and Aquifer Storage$150-2000/AF (highly variable)36-60 monthsVariableLowPumps for injection wells would require a moderate amount of power, likely requiring coordination with Rocky Mountain Power.
Utah law restricts rainwater collection to 2,500 gallons per parcel and has complex permitting for aquifer recharge, making large-scale capture difficult. [38, 44, 48]
Funding Municipal Recycled Water InfrastructureVariable contribution24-48 monthsProportional to fundingMediumThe power infrastructure would be managed by the municipal utility as part of the larger project.
The local South Valley Water Reclamation Facility treats wastewater and neighboring South Jordan has a water reuse demonstration project, indicating potential partnership opportunities. [41, 50, 52]
Solar-Thermal Chiller$800M - $1.6B5-8 years900-1200 (hypothetical)LowWhile Utah has a strong solar resource, this project's viability is constrained by land availability, not power. [19, 25, 37]
Finding the required ~400 acres of contiguous land in suburban West Jordan is a major, likely insurmountable, obstacle. [3, 4, 9]
Aligned Data Centers West Jordan, UT Water-free / closed-loop

Cooling as reported: Closed-loop and waterless cooling systems

Water: West Jordan City · Colorado River sourced · 47.4 million gallons in one year · solar-thermal: ~160 ac needed

Water use (estimated): ~39–393 AF/yr (from ~234 MW × air-cooling water intensity; actual varies with efficiency & reuse)

Land for solar-thermal cooling: ~160 ac needed · 1,562 ac open (non-built) within 3 km, 4,068 ac already built — ample

Sewer authority: Jordan Basin Improvement District (regional wastewater conveyance/treatment authority serving West Jordan, UT; final treatment at Jordan Basin WRF / regional system) ref · blowdown records held here, obtainable by public-records request

Costed projects (capex · timeline · viability by local power & water):

ProjectCapexTimelineAF/yrViabilityPower / water at this location
Fund Local Water Recycling Expansion$10-30M2-5 years145MediumThe Jordan Valley/Jordan River wastewater system is already grid-connected and utility-scale treatment is feasible in West Jordan, but this is a utility-dependent expansion rather than a power-limited project; local solar is not required.
Yes—Jordan Valley Water Conservancy District and regional wastewater plants are real local counterparties with an existing reuse pathway in the Salt Lake Valley, but incrementally getting 145 AF/yr of Colorado River-equivalent benefit depends on available treatment capacity, conveyance, and where the recycled water actually displaces potable demand.
VWBA Agricultural Fallowing Credits$3-8M1-3 years145MediumPower is not a material constraint because this is a contractual offset project, not a physical energy project; the local grid/solar conditions do not drive feasibility.
Utah has real agricultural users and OpenET-based verification is workable, but finding enough willing fallowing acreage near the actual Colorado River basin accounting area is uncertain, so this is feasible only as a negotiated offset program rather than a guaranteed local water supply.
DPR Innovation Seed Funding$2-10M3-7 years145MediumWest Jordan is electrically suitable for a pilot-scale advanced treatment facility, but a DPR pilot is not constrained by generation; the main issue is utility/treatment integration and permitting.
Direct potable reuse is not an established full-scale pathway in Utah yet, but Utah has been moving on reuse policy and demonstration projects; the needed source water exists in municipal effluent, though converting it into credited Colorado River savings is a regulatory and accounting challenge.
On-Site Blackwater and Stormwater Recharge$20-50M3-6 years145LowThe local grid can serve an advanced treatment plant, but the project’s scale is dominated by treatment/recharge operations rather than power availability; solar is irrelevant.
This is the weakest fit locally because aquifer recharge via injected treated effluent is heavily regulated and site-specific in Utah, and West Jordan’s shallow groundwater/recharge permissions, monitoring, and water-quality approvals would make achieving credited Colorado River benefits difficult.
Provo River Watershed/River Restoration$5-20M1-4 years145LowNo meaningful power constraint exists because this is a restoration/payment project, not an energy-intensive one.
The Provo River is a real regional water source, but translating watershed restoration into defensible, additional Colorado River water benefits from West Jordan is indirect and highly attribution-dependent, so it is not a strong local offset mechanism compared with direct reuse or verified fallowing.
Solar-Thermal Chiller$320-640M4-8 years145LowUtah has good solar resource, but the site’s nearby land is only partial and suburban; a roughly 160-acre solar-thermal field for an ~80 MW load is a poor fit for the local land base and would be difficult to permit and site.
This does not solve a water-availability problem because the facility already uses closed-loop/waterless cooling; there is no meaningful local water source advantage, and the project is mainly limited by land, siting, and economics rather than water.
DataBank Bluffdale, UT Cooling not reported

Cooling as reported: Not specified, but the company's lower water use suggests efficient cooling methods

Water: Bluffdale City · Colorado River sourced · 7.7 million gallons over a 12-month period

Water use (estimated): ~77–463 AF/yr (from ~45.9 MW × unknown-cooling water intensity; actual varies with efficiency & reuse)

Land for solar-thermal cooling: ~92 ac needed · 5,550 ac open (non-built) within 3 km, 877 ac already built — ample

Sewer authority: South Valley Sewer District (SVSD) ref · blowdown records held here, obtainable by public-records request

Costed projects (capex · timeline · viability by local power & water):

ProjectCapexTimelineAF/yrViabilityPower / water at this location
Volumetric Water Benefit Creditsn/a< 3 months23.6HighThis project has no impact on facility power consumption or the local grid.
Verified water offset markets, such as Bonneville Environmental Foundation's Water Restoration Certificates, exist for the Colorado River Basin, offering a contractual and immediate path to zero net water. [18, 23, 24]
Recycled Water Infrastructure$1-3M contribution18-36 months23.6HighThe additional energy for pumping recycled water represents a minor, manageable load for the local grid.
The South Valley Water Reclamation Facility is located in Bluffdale and serves the area, making a 'purple pipe' connection geographically feasible under established state non-potable reuse regulations. [34, 38]
Waterless Cooling Upgrade$5-15M+24-48 months23.6MediumThis may increase electricity use, posing a challenge given that Utah's grid is strained by data center growth, leading to long interconnection queues and a turn toward private power generation. [9, 20, 32]
This project permanently eliminates the need for 23.6 acre-feet of external water for cooling, offering the highest on-site water certainty.
Direct Potable Reuse Sponsorship$2-5M+ contribution3-7 years23.6LowSponsoring a DPR facility would support an energy-intensive project, adding a significant future load to the strained local power grid. [9, 20]
This project is contingent on the finalization of statewide Direct Potable Reuse (DPR) regulations; the nearby Pure SoJo project is a demonstration facility helping to inform these future rules. [7, 16, 26]
On-site Stormwater Recharge$500,000 - $2M12-24 monthsVariable, unlikely to meet the full 23.6 AF target annuallyMediumThis gravity-fed project requires no operational power, avoiding any impact on the grid.
Success is dependent on Bluffdale's variable annual precipitation (approx. 14-19 inches) and may not reliably offset the facility's full water use every year. [2, 3, 6]
QTS Data Centers Eagle Mountain, UT Water-free / closed-loop

Cooling as reported: Closed-loop cooling system

Water: Eagle Mountain City · Colorado River sourced · The closed-loop system will only need to be filled once and is not expected to consume additional water for cooling once operational · solar-thermal: ~240 ac needed

Water use (estimated): ~30–297 AF/yr (from ~177 MW × air-cooling water intensity; actual varies with efficiency & reuse)

Sewer authority: Eagle Mountain City Public Works / wastewater collection; final regional treatment is through the Utah County service area that reports to the Utah Department of Environmental Quality (publicly accessible records did not clearly identify a separate on-site industrial POTW connection for this facility) · blowdown records held here, obtainable by public-records request

Costed projects (capex · timeline · viability by local power & water):

ProjectCapexTimelineAF/yrViabilityPower / water at this location
VWBA agricultural efficiency$1-5M per ~100-500 AF/yr offset; often structured as a payment rather than owned assets6-18 months0.5-2.0MediumNo meaningful power dependency, so local grid limits are not the constraint; Utah/UTS is a strong solar state but this project is mainly a finance/contracting exercise.
This is the most plausible local offset path because Utah has active agricultural conservation/transfer opportunities in the Colorado River Basin, but it requires a willing counterparty and verified consumptive-use reduction rather than a physical water source on-site.
On-site blackwater/greywater recycling$8-20M18-36 months0.02-0.10LowPower is not the main issue; Eagle Mountain can supply load, but any on-site treatment adds significant electrical and operational complexity to an already large data-center campus.
Eagle Mountain City and Utah reuse rules make internal reuse feasible in principle, but a data center has very little blackwater/greywater to recycle and closed-loop cooling already minimizes demand, so the actual recoverable volume is tiny.
Local watershed restoration sponsorship$0.5-3M3-12 months0.0LowNo power dependence, but also no direct link to load or cooling demand, so it does not solve a physical water balance issue.
Restoration in Utah can improve habitat, recharge, or stream conditions, but it generally does not count as a defendable Colorado River consumptive-use offset unless paired with a quantified, legally recognized water-right change.
Solar-thermal chiller$480-960M for ~240 acres of collectors at $2-4M/acre48-84 months0.0-0.1LowUtah's solar resource is strong, but a ~120 MW data-center cooling duty would require a very large collector field, thermal storage, and balancing infrastructure, and the site’s actual cooling demand is likely far below a full 120 MW thermal requirement because it uses closed-loop cooling.
There is no special local water source issue this solves; the problem is scale and practicality, not water availability, and the project would not create a meaningful Colorado River water offset.

Wyoming 1 site

Microsoft Cheyenne, WY Evaporative / water-cooled

Cooling as reported: Existing facilities use direct evaporative cooling for less than 10% of the year. Newer designs are moving towards eliminating the need for ongoing water for cooling after an initial fill

Water: Cheyenne Board of Public Utilities · Colorado River sourced · Not specified

Water use (estimated): ~2,100–3,780 AF/yr (from ~250 MW × evap-cooling water intensity; actual varies with efficiency & reuse)

Sewer authority: City of Cheyenne Board of Public Utilities ref · blowdown records held here, obtainable by public-records request

Costed projects (capex · timeline · viability by local power & water):

ProjectCapexTimelineAF/yrViabilityPower / water at this location
Recycled Water for Cooling$2-5M24-48 months200-400MediumThe local grid can support the additional pumping load, which is minor compared to data center demand.
Cheyenne has a water reclamation program, but a recent contamination incident from a data center contractor has suspended industrial wastewater acceptance, complicating the timeline. [31, 36, 37, 41]
Agricultural Water Offset$400-800/AF annual lease12-24 months200-400HighThis project has no direct power requirement for the data center.
Local partners like the Laramie County Conservation District exist to facilitate voluntary irrigation efficiency projects and water conservation with agricultural users. [26, 27, 28]
Brackish Groundwater Desalination$3-7M36-60 months200-400LowHigh energy demand for reverse osmosis would require significant dedicated power infrastructure, though new large-scale generation is being developed in the area. [32]
While deeper, poor-quality groundwater exists in Laramie County, its suitability is unconfirmed, and brine disposal presents a major environmental and regulatory hurdle. [33, 38]

Colorado Front Range 6 sites

QTS Data Centers Aurora, CO Water-free / closed-loop

Cooling as reported: Closed-Loop

Water: Aurora Water · Colorado River sourced · The two QTS centers in Aurora use a combined 300,000 gallons in their closed-loop systems; ongoing use is primarily for office needs and outdoor irrigation. The average data center in Aurora uses about 500,000 gallons pe · solar-thermal: ~240 ac needed

Water use (estimated): ~27–269 AF/yr (from ~160 MW × air-cooling water intensity; actual varies with efficiency & reuse)

Land for solar-thermal cooling: ~240 ac needed · 4,537 ac open (non-built) within 3 km, 1,310 ac already built — ample

EPA facility record: QTS DEN1 - PHASE 1 (NPDES COR418198) · ECHO report

Sewer authority: City of Aurora / Aurora Water (wastewater collected by the municipal sewer system and treated at the Metro Water Recovery regional POTW serving Aurora) ref · blowdown records held here, obtainable by public-records request

Costed projects (capex · timeline · viability by local power & water):

ProjectCapexTimelineAF/yrViabilityPower / water at this location
On-Site Irrigation System Conversion to Reclaimed Water$0.5-2M12-24 months5-20MediumPower is not the limiter here; Aurora/Front Range grid access is generally strong enough for pumps and controls, but this project is water-network limited, not electricity-limited.
Aurora Water has a reclaimed-water system in the city, but actual service availability depends on whether the site can tie into reclaimed mains; Colorado reuse rules allow nonpotable irrigation use, yet the hard part is nearby reclaimed infrastructure and connection capacity, not source legality.
Agricultural Water Lease through a South Platte Water Bank$1-5M12-36 months50-300MediumNo material power requirement beyond administration; the Front Range grid and Aurora service area do not constrain an offsite lease transaction.
This is feasible because the South Platte Basin has existing water-rights markets and banks, but the offset must come from a real consumptive-use reduction or temporary interruptible supply agreement that survives Colorado water-court and augmentation requirements.
Investment in a Local Aquifer Recharge Project$3-15M24-60 months20-200MediumPower availability is not a major issue; recharge wells or basins are modest loads compared with data-center operations and the local grid is not the binding constraint.
Aquifer recharge is legally possible in Colorado, but it requires a carefully permitted source of recharge water, accounting for return flows and injury analysis; the practical challenge is securing a legally reusable source rather than finding a place to inject it.
Corporate Partnership for Colorado River Headwaters Restoration$0.5-5M6-24 months5-100LowPower is irrelevant here; this is a financial offset project rather than a physical water-supply project.
This is the weakest fit for getting to zero net Colorado River water in Aurora because headwaters restoration is usually not a direct, quantifiable consumptive-use offset under Colorado water law and may not create a defensible, site-specific basin accounting benefit.
Funding for Stormwater Capture and Reuse for Irrigation$1-4M12-30 months10-60MediumLocal power availability is adequate for pumps and controls, and the project scale is small relative to the Aurora-area grid.
Stormwater capture is workable in Aurora, but the dependable yield is limited and seasonal; it can reduce potable irrigation demand, yet it is not a large or fully reliable substitute unless paired with storage and permitted reuse infrastructure.
Solar-Thermal Chiller$480-960M48-96 months0LowAurora has good solar insolation by Colorado standards, but a 120 MW data center would need roughly 240 acres of collectors and large thermal storage; the land may exist on the east edge, yet this is still a very large, complex, intermittency-prone cooling build with significant integration risk.
This does not create a Colorado River water offset and does not solve the site's water accounting by itself; also, Colorado's dry, high-altitude climate helps solar resource but does not make solar-thermal cooling a practical substitute for conventional chilled-water systems at this scale.
CoreSite (DE1 & DE2) Denver, CO Water-free / closed-loop

Cooling as reported: Primarily air-cooled with some liquid cooling capabilities.

Water: Denver Water · Colorado River sourced · Phase 1 of CoreSite's new DE3 campus is estimated to use 235,000 gallons per day, significantly more than other Denver data centers which use roughly 10,000 gallons a day.

Water use (estimated): ~3–30 AF/yr (from ~18 MW × air-cooling water intensity; actual varies with efficiency & reuse)

Land for solar-thermal cooling: ~36 ac needed · 442 ac open (non-built) within 3 km, 5,667 ac already built — ample

EPA facility record: CORESITE - DE1 (Air permit CO0000000803102371) · ECHO report

Sewer authority: Metro Water Recovery (formerly the Metro Wastewater Reclamation District), which manages the Industrial Pretreatment Program for the City and County of Denver's Wastewater Management Division. ref · blowdown records held here, obtainable by public-records request

Costed projects (capex · timeline · viability by local power & water):

ProjectCapexTimelineAF/yrViabilityPower / water at this location
On-Site Water Treatment and Reuse$1.8M - $2.5M24-36 months263HighThe energy required for advanced treatment is manageable within the existing grid, especially with Denver's high solar potential which could support onsite generation. [4, 12, 29]
Colorado's 2022 Direct Potable Reuse (DPR) rule allows for this, and on-site wastewater from cooling tower blowdown provides a consistent source. [2, 3, 5, 6, 8, 39]
Fund Recycled Water Infrastructure ExtensionHighly variable; dependent on distance to Denver Water's recycled water lines.36-60 months263MediumPower for pumping is required, but the local grid, which is undergoing significant expansion, can support it. [16, 24, 41]
Denver Water has a recycled water program, but availability depends on the proximity of existing infrastructure to the data center sites. [5]
Aquifer Storage and Recovery (ASR) Partnership$7M+ for new wells, or partnership with an existing system.24-48 months for new build; less for partnership.263HighPumping is energy-intensive, but ASR is a priority for regional water providers, and grid upgrades are planned to accommodate new loads. [16, 30, 46]
The Denver Basin has vast storage potential, and entities like Denver Water and South Metro Water are actively developing ASR projects. [19, 21, 32, 43, 47]
Purchase Water Credits from Agricultural Conservation$52,000 - $85,000 per AF (one-time for permanent right)6-18 months263MediumThis project has no direct power requirement, making it independent of grid capacity.
Active water markets exist for Colorado Front Range water rights, though prices are high and reflect a permanent transfer from agriculture. [10, 18, 34, 49]
Direct Funding of South Platte River RestorationContribution to a larger project fund (e.g., WATR Program at $550M+ total). [9, 13]Ongoing263 (as an offset)LowThis project has no direct power requirement, making it independent of grid capacity.
While major restoration projects are underway, quantifying a direct acre-foot-for-acre-foot offset for data center use is legally and technically complex. [7, 9, 11, 17, 20]
CoreSite (DE3) Denver, CO Evaporative / water-cooled

Cooling as reported: Water-cooled chillers and cooling towers.

Water: Denver Water · Colorado River sourced · Phase 1 is estimated at approximately 235,000 gallons of water per day.

Water use (estimated): ~151–272 AF/yr (from ~18 MW × evap-cooling water intensity; actual varies with efficiency & reuse)

EPA facility record: CORESITE REAL ESTATE DE3 - CORESITE DE3 (NPDES COR423465, Air permit CO0000000803102609) · ECHO report

Sewer authority: Metro Wastewater Reclamation District (Denver metro regional wastewater utility; also serving the Denver Tech Center area where CoreSite DE3 is located) · blowdown records held here, obtainable by public-records request

Costed projects (capex · timeline · viability by local power & water):

ProjectCapexTimelineAF/yrViabilityPower / water at this location
Recycled Water Cooling Conversion$15-35M24-48 months77-95MediumXcel Energy’s Denver load zone is generally serviceable for a retrofit of this size, but any new pumps/treatment/controls add modest load; Denver has strong solar potential but onsite solar cannot materially offset cooling electricity on a 24/7 basis.
Denver has treated wastewater reuse activity, but a true on-site/reliably firm recycled-water supply for DE3 is not guaranteed at this address, and Colorado reuse rules still require an actual contractable reclaimed-water source and utility delivery infrastructure; potable-to-recycled conversion is more a distribution/partnership problem than a technology problem.
Watershed Restoration Investment$2-10M12-36 months10-50LowPower availability is not the binding constraint here; this is a financial offset project with negligible grid dependence beyond program administration.
In the Colorado Front Range, watershed and streamflow restoration can be legitimate mitigation, but converting it into zero-net Colorado River water is highly uncertain because the benefit is indirect, seasonal, and hard to quantify against Colorado River accounting; there is no local physical water source created, only an offset claim.
Agricultural Fallowing Water Credits$0.5-3M6-24 months25-100MediumPower is irrelevant except for program administration; the local grid can easily support this option.
Colorado does have agricultural temporary-fallowing and leasing concepts, but a Denver-based data center needs a real, legally transferable, Colorado-River-basin-suitable counterparty and basin accounting to claim net zero; the main constraint is not water availability but finding willing irrigators and getting durable, legally recognized credits.
Closed-Loop/Hybrid Cooling Retrofit$8-25M18-36 months60-110HighThis is the most grid-friendly option because it reduces water pumping and tower makeup while only modestly changing electrical load; Denver’s grid and solar resource can support the retrofit, though the facility will still rely on utility power for IT and cooling.
This does not require a new water source; it directly reduces consumption from the current water-cooled chiller/cooling-tower setup, which is practical in Denver where potable water is available but costly to justify for large evaporative loads.
Direct Potable Reuse (DPR) Advancement$20-60M36-72 months77-95LowXcel can serve the electrical side, but DPR adds treatment energy and operational complexity; solar resource is good, yet it does not solve the need for continuous high-reliability power and advanced treatment redundancy.
Colorado’s current potable reuse framework is still evolving and DPR is highly utility- and permit-dependent; Denver has reuse experience, but a data-center-specific DPR supply for DE3 is not a realistic near-term dedicated source without a municipal partner, plant investment, and regulatory approvals.
Novva Data Centers Colorado Springs, CO Water-free / closed-loop

Cooling as reported: Air-Cooled / Water-Free Cooling

Water: Colorado Springs Utilities · Colorado River sourced · **FLAGGED: AIR-COOLED/WATER-POSITIVE.** Novva states their proprietary system results in "almost zero water" usage, saving millions of gallons annually compared to water-cooled designs. · solar-thermal: ~400 ac needed

Water use (estimated): ~7–67 AF/yr (from ~40 MW × air-cooling water intensity; actual varies with efficiency & reuse)

Land for solar-thermal cooling: ~400 ac needed · 2,672 ac open (non-built) within 3 km, 627 ac already built — ample

Sewer authority: Colorado Springs Utilities · blowdown records held here, obtainable by public-records request

Costed projects (capex · timeline · viability by local power & water):

ProjectCapexTimelineAF/yrViabilityPower / water at this location
Solar-Thermal Chiller$800M - $1.6B4-7 years0 (potentially negative)LowThe project requires significant land and capital but does not generate electricity; it consumes it for pumps and fans, while the primary 200 MW data center load remains on the grid. [23, 25]
This project is counterproductive, as absorption chillers require a wet cooling tower for heat rejection, which would introduce significant water consumption to a facility currently using water-free cooling. [11, 37]
Water Rights Acquisition (Instream Flow)$52,000 - $85,000 per AF/yr (one-time)12-24 months20 (scalable, based on assumed non-cooling water use)MediumThis project has no impact on the data center's power consumption or the local grid.
Water rights in the Colorado Front Range are extremely expensive and transactions are complex, but it is a well-established market for permanently offsetting consumption. [29, 46, 47]
Agricultural Efficiency Offset$140 - $510 per AF/yr (annual or one-time funding)6-18 months20 (scalable, based on assumed non-cooling water use)HighThis project has no impact on the data center's power consumption or the local grid.
Funding agricultural efficiency is a cost-effective and established method for water conservation in Colorado, with programs paying farmers to fallow land or upgrade irrigation. [44, 45]
Municipal Recycled Water PartnershipVariable (e.g., $500k - $5M+ to fund specific infrastructure)24-48 months20 (scalable, based on assumed non-cooling water use)MediumThis project has no direct impact on the data center's power needs, but water recycling facilities are energy-intensive for the utility.
Colorado Springs Utilities is actively developing direct potable reuse (DPR) and has a large non-potable system, creating partnership opportunities to fund expansion in exchange for offsets. [13, 14, 18, 20]
EdgeConneX (DEN01 & DEN02) Englewood/Centennial, CO Water-free / closed-loop

Cooling as reported: N+1 CRACs, forced air plenum; likely air-cooled.

Water: Denver Water · Colorado River sourced · Not reported.

Water use (estimated): ~5–45 AF/yr (from ~27 MW × air-cooling water intensity; actual varies with efficiency & reuse)

Land for solar-thermal cooling: ~54 ac needed · 3,613 ac open (non-built) within 3 km, 2,769 ac already built — ample

Sewer authority: Arapahoe County Water and Wastewater Authority (ACWWA) appears to be the local wastewater/sewer utility serving the Englewood/Centennial area near this site; I could not verify a site-specific sewer connection from public records alone. · blowdown records held here, obtainable by public-records request

Costed projects (capex · timeline · viability by local power & water):

ProjectCapexTimelineAF/yrViabilityPower / water at this location
Off-Site Ag Water Conservation (VWBA credits)$0.5-3M upfront; typically $800-2,500/AF of verified credit plus transaction/MRV costs6-18 months10-100 AF/yrMediumPower is not the limiting factor here: the Front Range grid can support an off-site credit strategy because it is mostly a financial/administrative procurement, not an interconnection-heavy load. Local grid constraints only matter if the project also adds major on-site pumps or treatment.
Colorado has real agricultural consumptive-use reduction opportunities, but any VWBA-style offset must be durable and legally transferable; in Colorado that means complex water-rights administration and depletions accounting, not a simple purchase. This is feasible in principle on the Front Range, but it is not a guaranteed easy offset because legal water banking/administration and river-basin-specific accounting are tight.
Local Aquifer Storage and Recovery (ASR)$3-15M+ depending on wells, treatment, pilot testing, and permitting2-5 years0-50 AF/yrLowPower is generally available for pumping and controls, but ASR itself does not hinge on grid capacity; the bigger issue is whether a viable aquifer and water quality compatibility exist. Solar is ample on the Colorado Front Range, but it does not solve hydrogeologic or permitting barriers.
ASR is possible in Colorado, but only where a suitable aquifer, injection/recovery chemistry, and water-rights framework exist; Denver-area hydrogeology is highly site-specific and not every utility allows it as a simple mitigation tool. There is no broad, ready-made brackish resource in Englewood/Centennial that makes ASR a straightforward offset.
Landscaping Retrofit + Recycled Water for Irrigation$0.25-2M for xeriscape retrofit and purple-pipe connection, if nearby recycled supply exists12-36 months1-20 AF/yrMediumGrid capacity is not a major constraint; irrigation pumping and any small treatment/booster loads are modest. The region also has strong solar resource, but again the main limit is water-source proximity and utility service compatibility, not electricity.
Recycled water is real on the Front Range, but access is location-dependent: Denver Water has nonpotable and recycled-water programs, yet service lines are not universally available to every parcel in Englewood/Centennial. For a data center footprint, irrigation demand is usually too small to produce much Colorado River offset, so this is more a local demand-reduction measure than a large water-neutrality solution.
Headwaters Restoration Fund Contribution$0.2-2M donation/grant-style contribution; often modeled as $500-3,000/AF equivalent3-12 months5-50 AF/yr equivalent, depending on fund methodologyMediumPower availability is largely irrelevant because this is a financial contribution rather than a load-changing project. The Colorado Front Range grid and solar resource do not materially affect feasibility.
Headwaters/restoration funds exist in the West, but whether they count as credible Colorado River water neutrality depends on the fund’s accounting standard, permanence, and basin linkage. In Colorado, the hardest part is proving that ecological restoration translates into a real, defensible consumptive-use offset rather than a general environmental benefit.
100% Closed-Loop and On-Site Water Reuse$10-40M+ for cooling-system redesign, water treatment, storage, heat rejection, and possible chiller upgrades18-48 months10-80 AF/yrMediumThe Front Range grid can support the electrical load of higher-efficiency chillers, pumps, and water treatment, and the area has good solar resource if the owner wants to pair with on-site PV or PPAs. The bigger practical issue is that closed-loop systems often raise electric demand even as they cut water use, so interconnection and backup power sizing must be checked carefully.
This is the most technically direct way to reduce potable demand, but for an existing air-cooled/N+1 CRAC site it may be limited by the current thermal architecture and by winter freeze, blowdown, and water-quality constraints. Colorado does allow reuse pathways, but a truly closed-loop retrofit at an operating edge data center is usually capital-intensive and site-specific rather than a simple Denver Water service swap.
Google Thornton, CO Water-free / closed-loop

Cooling as reported: Google employs a variety of cooling methods, including water-based, air-cooling, and hybrid systems, often tailored to the specific location.

Water: City of Thornton (receives some water from Denver Water) · Colorado River sourced · Not reported for this specific site. Google's average data center consumes about 450,000 gallons per day.

Land for solar-thermal cooling: 2,142 ac open (non-built) within 3 km, 3,345 ac already built — n/a (MW unknown)

EPA facility record: GOOGLE, INC (Air permit CO0000000800101861) · ECHO report

Sewer authority: Metro Water Recovery · blowdown records held here, obtainable by public-records request

Costed projects (capex · timeline · viability by local power & water):

ProjectCapexTimelineAF/yrViabilityPower / water at this location
On-Site Recycled Water System for Cooling$1.5-6M24-36 months400-500HighThe additional power load is manageable for the local grid, and on-site solar could offset a portion of the energy required for treatment. [11, 24]
The City of Thornton has an established reclaimed water program (Regulation 84), and Google can treat its own process water for direct reuse in cooling systems. [2]
Aquifer Storage and Recovery (ASR) in the Denver Basin$7-15M36-60 months500+HighEnergy is required for injection and recovery pumps, but this load is flexible and can be scheduled during off-peak hours, which is compatible with Xcel's grid. [14, 15]
The data center sits above the Denver Basin aquifers, and ASR is a proven, legally supported strategy in the region for storing water. [10, 15, 18, 22]
Voluntary Water Benefit Agreement (VWBA) via Agricultural Fallowing$150-500/AF annually3-6 months to establish annual agreement500+MediumThis project has no direct power requirement, making it independent of grid constraints.
Finding willing agricultural partners in the South Platte Basin for temporary fallowing is feasible, but competition and water right complexities can pose challenges. [37, 41]
Funding for South Platte River RestorationContribution-based (e.g., $5-10M)Ongoing (project dependent)Indirect/modeled benefitMediumThis project has no direct power requirement for Google.
Large-scale restoration projects are underway on the South Platte, offering partnership opportunities, though quantifying direct acre-foot benefits to Google is complex. [38, 42, 43]
Advanced On-Site Water Generation and Reuse$5-10M+24-48 months10-500 (varies by tech)LowAtmospheric water generation is extremely energy-intensive, and while direct potable reuse is possible, its power demands are significant and may strain local grid resources without dedicated generation. [47, 48]
While Colorado approved direct potable reuse (DPR) in 2022, implementation requires extensive design and regulatory approval; atmospheric generation is low-yield for the cost in this semi-arid climate. [2, 3, 5, 7]

New Mexico (San Juan-Chama) 2 sites

Meta (Facebook) Los Lunas, NM Evaporative / water-cooled

Cooling as reported: Evaporative cooling, designed to be more water-efficient than industry standard.

Water: Village of Los Lunas / Albuquerque Bernalillo County Water Utility Authority (ABCWUA) · Colorado River sourced · **FLAGGED: WATER-POSITIVE (GOAL).** Consumed 75 million gallons in 2023. Meta has a goal to be water-positive by 2030 and invests in local water restoration projects that are projected to restore over 172 million gallons

Water use (estimated): ~1,008–1,815 AF/yr (from ~120 MW × evap-cooling water intensity; actual varies with efficiency & reuse)

Land for solar-thermal cooling: ~240 ac needed · 5,760 ac open (non-built) within 3 km, 1,134 ac already built — ample

Sewer authority: Village of Los Lunas wastewater/sewer system (local municipal wastewater utility); any regional treatment may also involve the Albuquerque Bernalillo County Water Utility Authority (ABCWUA) as the broader water/wastewater authority for the area, but the specific publicly confirmable sewer authority for this site is not clearly published in accessible primary sources. · blowdown records held here, obtainable by public-records request

Costed projects (capex · timeline · viability by local power & water):

ProjectCapexTimelineAF/yrViabilityPower / water at this location
Fund Municipal Recycled Water Infrastructure$25-60M (utility-side; ~$1,200-3,000/AF-yr for a 20-50 MGD recycling buildout share)3-7 years500-1,500MediumThe Albuquerque/Los Lunas load area is on the PNM grid with a strong solar resource, but recycled-water delivery is mostly a pumping/conveyance problem rather than a bulk-power problem; utility-side interconnection is straightforward compared with new data-center generation.
This is plausible locally because Los Lunas is served by the Village of Los Lunas/ABCWUA system and the region already has municipal wastewater infrastructure, but New Mexico reuse is governed by state discharge/reuse permitting and any direct nonpotable reuse needs a real treatment/distribution project, not just a contract. ABCWUA and the South Valley/San Juan-Chama context make reclaimed water more realistic than brackish supply, but available volumes are limited and not all effluent is legally or physically recoverable.
Convert to Hybrid Air-Cooled Systems$8-20M12-30 months600-1,200MediumLos Lunas has good solar and a normal utility grid, so the additional fan/chiller-electric load of hybrid cooling is workable; the bigger constraint is peak summer power quality and redundancy rather than energy availability.
This directly reduces dependence on potable supply, which is important because the local water system is not unlimited; however, it does not require a special local water source, only engineering changes, and in hot NM summers hybrid systems can still need some make-up water for extreme conditions.
Sponsor an Aquifer Storage and Recovery (ASR) Project$10-30M4-8 years300-1,000LowASR requires modest pumping power and the grid can support it, but the project is not power-limited; the issue is hydrogeology, permitting, and treatment reliability.
ASR is hard here because New Mexico ASR projects need a suitable aquifer, water right, and state approvals, and the Rio Grande basin is heavily regulated; there may be no easy, unused offset counterparty at scale near Los Lunas, and injected water quality/compatibility is a major constraint. This makes it much less certain than in states with mature ASR markets.
Lease Water Rights via Verified Agricultural Fallowing$1-5M (transaction, compensation, monitoring)1-3 years400-1,200MediumPower is not a binding constraint; this is a legal/market transaction rather than an energy-intensive project, so the local grid and solar resource are largely irrelevant.
This can work in the Middle Rio Grande if a real, verified water-rights holder is willing to fallow land and the transfer conserves consumptive use under New Mexico rules, but the market is thin and highly constrained by basin administration, acequia/community concerns, and interstate/rio Grande accounting. It is feasible in principle, but scale and timing are uncertain.
Develop On-Site Wastewater Reuse$6-18M18-36 months200-700HighThe local grid and solar resource can easily support tertiary treatment, UV, pumps, controls, and storage; on-site reuse is power-light compared with a new generation or transmission project.
This is the most site-controlled option and does not depend on finding an outside offset counterparty; New Mexico generally allows industrial water reuse through permitting and engineering controls, and on-site reuse can offset potable make-up demand directly. The limitation is quantity, because a data center's wastewater volume is usually much smaller than its cooling make-up need, so it helps but will not fully zero out water use by itself.
Project Jupiter (Oracle - Unconfirmed) Los Lunas, NM Water-free / closed-loop

Cooling as reported: "Closed-loop" system.

Water: Village of Los Lunas / ABCWUA · Colorado River sourced · A one-time fill of 11 million gallons is reported, with the system then recycling that water.

Water use (estimated): ~168–1,680 AF/yr (from ~1000 MW × air-cooling water intensity; actual varies with efficiency & reuse)

Reported: ~1 AF/yr company env report

Sewer authority: Village of Los Lunas Wastewater Division ref · blowdown records held here, obtainable by public-records request

Costed projects (capex · timeline · viability by local power & water):

ProjectCapexTimelineAF/yrViabilityPower / water at this location
Brackish Water Desalination$35-115M for a 1-5 MGD plant3-5 years for siting, permitting, and construction1,120 - 5,600MediumThe local grid has significant existing and planned solar capacity, some built specifically for large industrial users, making it capable of supporting the energy-intensive desalination process. [9, 37, 38]
Brackish water is available in the Albuquerque Basin, but site-specific exploration is required to confirm quantity and quality. [39, 40, 45]
Purchase & Retire Agri. Water Rights~$15,000/AF (based on 2012 data)12-24 months per transactionScalable based on investmentHighThis project has no significant power requirements.
An active, though informal, market exists for water rights in the Middle Rio Grande, with a history of ag-to-urban transfers. [13, 48]
Voluntary Agricultural Fallowing~$2M to establish a pilot program12-18 months to establish program and enroll participantsScalable based on enrollmentHighThis project has no significant power requirements.
Pilot programs are being actively studied in New Mexico, and the Middle Rio Grande Conservancy District has a large agricultural user base that could participate. [5, 35, 48]
Aquifer Storage and Recovery (ASR)$8-10M for a small to medium-scale project5-7 years for permitting and construction3,000+MediumPumping for injection and recovery is energy-intensive, but the local grid is robust and supported by significant renewable generation. [9, 17, 38]
The local aquifer is a candidate, but a key constraint is securing a source water for injection, such as treated effluent from the Village of Los Lunas or ABCWUA. [3, 14, 17, 25]
Rio Grande Water Fund Contribution~$700 per acre of watershed restoredImmediate (contribution); 20-year restoration planIndirect benefit; improves long-term yield and resilience of the Rio GrandeHighThis project has no significant power requirements.
The fund's upstream restoration work is designed to directly enhance the quantity and quality of the water source for Los Lunas and the entire Middle Rio Grande region. [6, 10, 21]

California Inland Empire 6 sites

One Data Center America (ONT-1 & ONT-2) Ontario, CA Cooling not reported

Cooling as reported: Not specified, but notes "cooling redundancy."

Water: City of Ontario (via Inland Empire Utilities Agency, a member of MWD) · Colorado River sourced · Not reported.

Water use (estimated): ~2–11 AF/yr (from ~1.1 MW × unknown-cooling water intensity; actual varies with efficiency & reuse)

Land for solar-thermal cooling: ~2 ac needed · 1,071 ac open (non-built) within 3 km, 5,359 ac already built — ample

Sewer authority: City of Ontario Municipal Utilities Company (wastewater collection) with regional treatment by Inland Empire Utilities Agency (IEUA) at the regional water reclamation plant level ref · blowdown records held here, obtainable by public-records request

Costed projects (capex · timeline · viability by local power & water):

ProjectCapexTimelineAF/yrViabilityPower / water at this location
Cooling System Retrofit to Eliminate Water Use$20-60M+24-48 months0.8-1.8LowOntario is in a strong-grid, high-solar-resource part of the Inland Empire, but a true zero-water retrofit usually means air-cooled or hybrid-to-dry cooling that materially increases power demand and peak load; SCE transmission/service exists, but large incremental electric load can still face utility interconnection and substation lead times.
This is technically available anywhere, but for a hyperscale-style data center in hot Inland Empire conditions it is the least practical path to zero Colorado River water because it trades water for major energy use and efficiency penalties; it does not depend on local recycled-water rules or offset counterparty availability.
Switch to 100% Recycled Water for Cooling Needs$2-8M12-36 months0.8-1.8HighPower availability is favorable for keeping water-cooled systems in place because this option avoids large added electrical load; the local grid is robust for the Inland Empire and solar resource is excellent for onsite/offsite renewable support, though that does not remove water-supply permitting needs.
This area has real recycled-water infrastructure through Inland Empire Utilities Agency and regional wastewater reuse in the Chino Basin/Western Riverside basin, but whether Ontario/IEUA can deliver reliable nonpotable recycled water to the site depends on proximity, pipeline connection, pressure, and contract capacity; California recycled-water use is permitted and encouraged, but service is location-specific.
Investment in the Chino Basin Groundwater Recharge Program$1-10M6-18 months0.5-2.0MediumThis is not power-intensive and therefore fits the local grid easily; Ontario/Inland Empire has strong utility infrastructure and solar resource, so the limiting factor is not electricity but whether a recharge/offset structure can be contracted and verified.
The Chino Basin is a real managed groundwater basin with active recharge and groundwater management, but using recharge as a direct Colorado River water offset requires a legally recognized accounting pathway and a willing local program sponsor; it is feasible in this basin, yet crediting is not the same as a guaranteed on-site water supply.
On-Site Stormwater Capture and Reuse System$1-5M9-24 months0.05-0.3LowPower is not the constraint; pumps and treatment are small loads relative to a data center, and the regional grid/solar resource are adequate.
Ontario can physically capture stormwater, but Inland Empire rainfall is limited and highly seasonal, so on-site runoff volume is usually too small and too variable to materially offset cooling demand; California allows stormwater capture and reuse, but the available yield here is modest.
Purchase of Voluntary Water Benefit Agreement (VWBA) Credits$0.1-1M3-12 months0.5-2.0MediumPower availability is largely irrelevant because this is a market/contract instrument rather than an energy project; the local grid and solar resource do not constrain credit purchase, only the ability to verify and retire credits.
VWBA-style credits can exist where watershed projects generate quantified benefits, but the Inland Empire is not a simple one-to-one Colorado River offset market and credits must be sourced from real, additional, and properly verified projects; availability is possible but counterparty quality and additionality are the main risks.
Iron Mountain Riverside, CA Cooling not reported

Cooling as reported: Not specified

Water: Riverside Public Utilities (receives some MWD water) · Colorado River sourced · Not reported. Iron Mountain has a portfolio of data centers in the US.

Land for solar-thermal cooling: 4,973 ac open (non-built) within 3 km, 1,307 ac already built — n/a (MW unknown)

Sewer authority: City of Riverside, Public Works Department, Wastewater Division · blowdown records held here, obtainable by public-records request

Costed projects (capex · timeline · viability by local power & water):

ProjectCapexTimelineAF/yrViabilityPower / water at this location
Recycled Water for Cooling$0.5-1.5M12-24 months100 (assumed)HighThis project has a negligible impact on power consumption, which the local grid can easily support. [40]
Riverside has an established recycled water program suitable for industrial cooling, making the source viable assuming proximity to distribution lines. [8, 47, 48]
Agricultural Fallowing Water Offsetn/a3-6 months100 (assumed)HighThis contractual project has no impact on the data center's power requirements or the local grid.
Metropolitan Water District and other entities facilitate water transfer and offset programs, making a counterparty in a nearby irrigation district plausible. [42]
On-site Liquid Cooling Retrofit$2-5M18-36 months100 (assumed)HighThe improved energy efficiency of liquid cooling reduces the overall load, which is a benefit to the grid being upgraded by the Riverside Transmission Reliability Project. [15, 40]
This project creates a closed-loop system that virtually eliminates the need for an ongoing external water source for cooling. [19]
Brackish Groundwater Desalination$80,000-250,00024-48 months100 (assumed)MediumThe small energy load for a plant of this size is easily supported by the local grid and the region's high potential for on-site solar generation. [4, 29]
The Inland Empire has known brackish groundwater aquifers and the Inland Empire Brine Line is available for concentrate disposal, making the project technically feasible. [13, 23, 31]
Lumen (formerly Level 3 / tw telecom) Ontario, CA Cooling not reported

Cooling as reported: Not specified

Water: City of Ontario (via IEUA, a member of MWD) · Colorado River sourced · Not reported. This is a legacy telecom and colocation site.

Land for solar-thermal cooling: ~1 ac needed · 1,164 ac open (non-built) within 3 km, 5,289 ac already built — ample

Sewer authority: City of Ontario Sewer Division / City of Ontario Public Works (local sewer collection), with wastewater treatment provided regionally through the Inland Empire Utilities Agency (IEUA) system ref · blowdown records held here, obtainable by public-records request

Costed projects (capex · timeline · viability by local power & water):

ProjectCapexTimelineAF/yrViabilityPower / water at this location
Convert to Recycled Water for Cooling$3-12M18-36 months20-120MediumOntario/IEUA sits in a strong Southern California grid area with ample solar resource, but recycled-water cooling still depends on existing utility interties, pumping, treatment, and cooling-tower controls more than on raw grid capacity.
This is the most location-fit offset because Inland Empire Utilities Agency supplies recycled water in the Ontario area, but actual delivery depends on pipe access, pressure, and water-quality compatibility for cooling systems.
Fund Local Groundwater Recharge$0.5-5M12-48 months5-50LowPower is not the main constraint; the area can support pumping and monitoring loads easily, but that does not solve the harder physical and regulatory recharge-facility constraint.
Groundwater recharge in this basin is real and managed locally, but a private data-center-funded offset is only credible if it is tied to an approved recharge project with measurable accounting; there is no simple one-to-one onsite depletion replacement.
Purchase Agricultural Water Conservation Credits$50k-500k6-24 months10-100LowPower availability in the Inland Empire does not materially help this project; the binding issue is whether a real, bankable counterparty exists.
Agricultural conservation credits are not a standard, liquid offset market in Ontario/IEUA/MWD for zero-net accounting, so any credit would need a bespoke, legally durable arrangement rather than an ordinary local purchase.
Upgrade to Waterless Cooling Technology$8-30M12-30 months20-150MediumThe site’s Southern California grid access and strong solar resource can support higher electric load from air-cooled or liquid/immersion systems, though peak-demand charges and utility interconnection study could be material.
This avoids the need for local substitute water entirely, which is attractive because the Inland Empire does have recycled water but not unlimited low-cost supply at every site; the tradeoff is much higher electricity use.
Sponsor Local Watershed Restoration$100k-2M6-36 months1-20LowPower is essentially irrelevant here; the question is whether restoration actions can be quantified as a durable water offset, which is difficult.
Watershed restoration can improve local hydrology and habitat, but it is generally not accepted as a direct acre-foot-for-acre-foot replacement for a data center’s consumptive use in California water accounting.
Aecero (CBX) Corona, CA (Riverside County) Cooling not reported

Cooling as reported: Not specified

Water: City of Corona (receives some MWD water) · Colorado River sourced · Not reported.

Water use (estimated): ~10–60 AF/yr (from ~6 MW × unknown-cooling water intensity; actual varies with efficiency & reuse)

Land for solar-thermal cooling: ~12 ac needed · 352 ac open (non-built) within 3 km, 6,228 ac already built — ample

Sewer authority: City of Corona Utilities Department ref · blowdown records held here, obtainable by public-records request

Costed projects (capex · timeline · viability by local power & water):

ProjectCapexTimelineAF/yrViabilityPower / water at this location
Air Cooling Retrofit$5,000-$15,000 per rack12-24 months150MediumAir cooling is more energy-intensive, increasing load on a Southern California Edison grid already facing a 5 GW data center queue, though the area has excellent solar resources. [50, 8, 19]
This project eliminates the need for on-site water for cooling, directly addressing the goal of zero water use. [51]
On-site Recycled Water Plant$1-5M24-48 months150HighThe project's power needs are moderate and can be met by the local grid or supplemented with on-site solar, which is highly viable in Corona. [8, 13, 39]
The City of Corona has an extensive reclaimed water system with 44 miles of purple pipe, making connection and supply feasible. [4, 7, 14]
Brackish Groundwater Desalination$350-$800/AF (production cost)36-60 months150MediumDesalination is highly energy-intensive, posing a significant new load on the SCE grid, which could be challenging given current interconnection queues. [16, 21, 50]
The nearby Chino Basin has active brackish water desalter authorities and projects, indicating potential for a local brackish source. [6, 11, 18, 20]
Agricultural Water Offsetn/a6-12 months (agreement negotiation)150HighThis project has no direct impact on the data center's power consumption or the local grid. [37]
Paying for irrigation efficiency in nearby agricultural districts like those in Imperial or Riverside Counties is a common water transfer mechanism to offset Colorado River water use. [38, 44]
Digital Realty (Planned) Vernon, CA (near Inland Empire) Cooling not reported

Cooling as reported: Not specified

Water: Vernon Public Utilities · Colorado River sourced · Not applicable (planned).

Water use (estimated): ~54–323 AF/yr (from ~32 MW × unknown-cooling water intensity; actual varies with efficiency & reuse)

Land for solar-thermal cooling: ~64 ac needed · 200 ac open (non-built) within 3 km, 6,564 ac already built — ample

Sewer authority: City of Vernon Public Works / Vernon Public Utilities sewer system (Vernon, CA); likely ultimately conveyed/treated under the regional Los Angeles County sanitation/wastewater system, but a site-specific industrial discharge connection/pretreatment record was not found in public sources ref · blowdown records held here, obtainable by public-records request

Five projects to zero out its river draw:

  1. One-line description
  2. Rough AF/yr avoided or returned
  3. Rough cost/structure
  4. Why it fits THIS site
  5. One-line description
Imperial Valley Computer Manufacturing (Proposed) Imperial, CA (Imperial County) Cooling not reported

Cooling as reported: Water-intensive cooling.

Water: Imperial Irrigation District (IID) · Colorado River sourced · Seeking ~287 million gallons per year (750,000 gallons/day) from the Colorado River after plans to use recycled water failed. This project is currently in litigation over water access. · solar-thermal: ~660 ac needed

Water use (estimated): ~554–3,327 AF/yr (from ~330 MW × unknown-cooling water intensity; actual varies with efficiency & reuse)

Reported: ~880 AF/yr government

Land for solar-thermal cooling: ~660 ac needed · 1,583 ac open (non-built) within 3 km, 957 ac already built — feasible

Sewer authority: City of Imperial, Department of Public Services ref · blowdown records held here, obtainable by public-records request

Costed projects (capex · timeline · viability by local power & water):

ProjectCapexTimelineAF/yrViabilityPower / water at this location
On-Site Brackish Water Desalination$8M-$18M24-48 months881MediumThe IID grid has capacity for new projects, but large loads (>20MW) require special tariffs and funding for infrastructure upgrades. [25, 36]
Brackish groundwater is available in Imperial Valley, but its specific location and suitability for this project would require further hydrogeological study.
Agricultural Fallowing Offset Program$264,300-$378,830 annually3-6 months881HighThis project has no significant power requirements.
The Imperial Irrigation District has multiple active, voluntary fallowing and conservation programs with established payment rates, recently ranging from $300 to $430 per acre-foot. [17, 19, 21]
Investment in Municipal Recycled Water Infrastructure$20M-$50M+3-5 years881LowAdvanced water treatment is energy-intensive and would add a significant load to the IID grid, likely requiring system upgrades. [7]
While nearby El Centro has a wastewater treatment plant, it is not currently set up for recycled water distribution, and new regulations for direct potable reuse are extensive and costly to implement. [4, 42, 49]
Advanced Hybrid Cooling System Conversion$15M-$30M18-24 monthsup to 790HighWhile potentially increasing the facility's power demand, the IID grid is actively upgrading and can accommodate new, large loads with proper planning and tariffs. [25, 37]
This project reduces water demand by up to 90% rather than requiring a new source, making it a highly viable internal solution.
Colorado River Tributary Restoration PartnershipHighly variable, likely $500-$2000/AF2-5+ years881LowThis project has no power requirements.
While restoration projects exist, quantifying and transferring the specific water savings to an out-of-state data center is a complex legal and administrative process with no clear precedent. [14, 18, 22]
Solar-Thermal Chiller$1.3B-$2.6B4-6 years881LowThe project generates its own thermal energy for cooling, reducing electrical grid load, and Imperial County has abundant solar resources and land designated for renewable energy. [23, 39]
This technology drastically reduces water consumption for cooling, but the extremely high capital cost and large land requirement make it economically unviable for this specific application.

← Back to the playbook · Sources and methods