Data Centers Eye Produced Water: 12-Hr Storage Rule [2026]

Data center operators scrambling for new water sources now have oilfield wastewater on their checklist. A report published by TechRadar on what data centers need to consider when storing produced water puts a spotlight on a question the industry has mostly avoided until now: can the high-salinity byproduct of oil and gas drilling actually cool the servers running the AI boom. The answer, according to tank manufacturers, water-reuse guidance documents, and infrastructure standards bodies, is a cautious yes, but only with storage, treatment, and monitoring requirements that most facilities are not yet built for.

The timing matters. AI-driven data center construction has pushed freshwater demand in drought-prone regions like the Permian Basin to the point where utilities, regulators, and site selectors are actively hunting for alternative supply. Produced water, generated in enormous volumes alongside every barrel of oil pumped from the ground, is one of the few sources available at the scale hyperscalers need. Storing it safely is where the real engineering work begins.

Google · Preferred Sources

Don't miss new tech stories on Google

Add Tech Insider once in the Google app and our stories appear in your news suggestions.

Add Now

What Produced Water Means for the Data Center Industry

Produced water is described by industry sources as high-salinity wastewater generated during oil and gas operations, a byproduct that has traditionally been reinjected underground or disposed of at considerable cost to drillers. For data centers, it represents something different: a non-potable water stream that does not compete with municipal drinking supply, agricultural irrigation, or residential use. That distinction is becoming central to how new data center campuses get sited, permitted, and defended in front of local water boards already wary of AI infrastructure’s thirst.

Tank manufacturer TanksCo lists produced-water storage explicitly among the applications for its bolted-steel tank systems, a sign that the demand side of this market is no longer theoretical. The company’s product line is built to NFPA 22 fire-protection standards and spans a wide capacity range, which matters because produced-water volumes at a single well pad or gathering station can vary from a few thousand gallons a day to industrial-scale flows that need permanent, large-footprint storage.

Inside the Report on Storing Produced Water at Data Centers

The TechRadar report frames produced-water storage as an emerging consideration rather than a settled practice, and that framing lines up with what the supporting technical documentation actually shows. There is no single, universally adopted produced-water storage code written specifically for data centers today. Instead, operators are stitching together requirements from fire-protection codes, Uptime Institute reliability tiers, and water-reuse quality guidance that was originally developed for reclaimed and recycled water more broadly, then applying it to this new use case.

That patchwork approach is workable in the short term, but it leaves gaps. A storage tank engineered to hold produced water safely from a corrosion and fire-safety standpoint is not automatically a tank that delivers water clean enough for direct or evaporative cooling equipment. Those are two separate engineering problems, and the report’s framing suggests the industry is only now starting to treat them as such.

Why Data Centers Are Looking at Oilfield Wastewater Now

The push toward produced water is not happening in isolation. Big tech’s data center buildout has already strained local water systems, as detailed in coverage of Google’s $15 billion India AI hub, which drew local backlash over water allocation before the company moved to defend the project. Similar tension is playing out across drought-exposed US regions, where utilities and grid operators are also managing the broader $1.4 trillion grid overhaul tied to AI power demand. Water and power are converging into the same siting conversation, and produced water offers a way to decouple at least one of those constraints from the public supply.

The Permian Basin Connection

West Texas sits at the center of this shift for a simple reason: it already has both abundant natural gas for onsite power generation and some of the highest produced-water volumes in the country from decades of Permian Basin drilling. Research cited by the University of Texas has examined the beneficial-use potential of produced water in the Permian specifically, treating it as a resource rather than a disposal liability. That reframing is what makes data center operators willing to consider it at all.

Bolted-Steel Tanks: How TanksCo and CST Industries Build for Produced Water

Storage infrastructure for produced water at data center sites borrows heavily from designs already proven in the oil and gas sector. TanksCo’s bolted-steel systems, built to NFPA 22 standards, range from 15,000 gallons up to more than 1,000,000 gallons, a span wide enough to cover everything from a single edge facility to a hyperscale campus with multiple cooling towers.

CST Industries approaches the same problem with a different materials strategy, offering its Aquastore glass-fused-to-steel tanks alongside the TecTank and OptiDome product lines for data center water storage. Glass-fused-to-steel construction is prized in water utility applications for corrosion resistance, a relevant feature given that produced water’s high salinity is exactly the kind of chemistry that degrades untreated steel over time.

The practical takeaway for operators evaluating vendors: tank material selection is not a commodity decision once produced water enters the picture. Corrosion allowances, coating specifications, and inspection intervals all need to be rated for high-salinity service, not just generic industrial water storage.

The Uptime Institute’s 12-Hour Rule and What It Means for Reliability Tiers

Water storage capacity at a data center is not just a sustainability metric, it is a reliability requirement. According to ProTank, the Uptime Institute Tier Standard calls for roughly 12 hours of onsite makeup-water storage at facilities that rely on evaporative cooling. That reserve exists so a facility can keep cooling systems running through a supply interruption, whether that is a municipal main break, a treatment plant outage, or in this case, a disruption to a produced-water delivery pipeline.

Tier III’s Concurrently Maintainable Requirement

ProTank’s guidance also notes that Tier III certification raises the bar further: storage must be concurrently maintainable, meaning a tank or section of the water system can be taken offline for maintenance without interrupting the overall reserve or dropping capacity below the required threshold. For produced-water systems, that typically means redundant tankage and treatment trains rather than a single large tank, since any one component may need isolation for cleaning, inspection, or repair without compromising uptime.

The 400,000-Gallon Math Behind Evaporative Cooling

ProTank’s figures also put a number on just how much water evaporative cooling actually consumes: roughly 400,000 gallons per megawatt of IT load annually. Scaled across a modern AI campus running into the hundreds of megawatts, that figure explains why operators are no longer willing to rely on a single water source, and why produced water, treated reclaimed water, and municipal supply are increasingly discussed as a blended portfolio rather than a single pipeline.

That consumption math is also why cooling technology choices matter as much as water sourcing. Facilities leaning harder on liquid and immersion cooling, the subject of ongoing investment covered in reporting on Trane Technologies’ AI cooling bet, reduce the sheer volume of water that needs to be sourced, stored, and treated in the first place, which in turn reduces how much produced water a site would ever need to handle.

Water Quality Benchmarks Before Reuse Can Happen

Storage is only half the engineering challenge. A water-reuse guidance document presented in a data center and reuse webinar hosted by the Association of California Water Agencies sets out specific treatment and quality thresholds that reclaimed water must meet before it reaches cooling infrastructure, and those benchmarks are the clearest technical signal available for what produced water would also need to clear.

Chemical and Disinfection Thresholds

The guidance calls for secondary wastewater treatment and filtration as a baseline, with a 0.5 mg/L disinfectant residual maintained in the pipeline between the recycle facility and the data center, and a minimum 0.2 mg/L free-chlorine residual inside direct evaporative-cooling units themselves. Turbidity has to stay below 3 NTU, must not exceed 5 NTU in 95% of measurements, and total suspended solids need to stay under 5 mg/L.

Biological Risk: Legionella and E. coli Limits

Biological contamination is treated just as strictly. E. coli should be nondetectable in 75% of samples, and no single sample can exceed 25 CFU per 100 mL. Legionella pneumophila, the bacteria behind Legionnaires’ disease and a known risk in cooling towers and evaporative systems generally, has to stay below 100 CFU/mL in both the storage tank and the direct evaporative-cooling units. Those thresholds exist because evaporative cooling towers aerosolize water, and any pathogen present in storage has a direct path into the air around the facility.

Water Quality MetricRequired ThresholdWhere It Applies
TurbidityBelow 3 NTU (max 5 NTU in 95% of samples)Treated water prior to cooling use
Total suspended solidsBelow 5 mg/LTreated water prior to cooling use
E. coliNondetectable in 75% of samples; no sample over 25 CFU/100 mLTreated water prior to cooling use
Legionella pneumophilaBelow 100 CFU/mLStorage tank and direct evaporative-cooling units
Disinfectant residual0.5 mg/LPipeline between recycle facility and data center
Free chlorine residual0.2 mg/L minimumDirect evaporative-cooling units

Comparing Produced Water to Municipal, Reclaimed, and Rainwater Supply

Every non-potable water source available to a data center carries a different cost and risk profile. Municipal supply is the easiest to treat but the most politically sensitive to draw down, especially in regions already fighting over residential allocation. Reclaimed municipal wastewater has an established regulatory path and the quality benchmarks described above, but availability depends on proximity to a treatment plant with spare capacity. Rainwater and stormwater capture is low-conflict but inconsistent and climate-dependent.

Produced water sits in its own category. Volumes can be enormous in active drilling regions, and it does not compete with any existing beneficial use, since it has historically been a disposal cost for operators rather than a resource. The tradeoff is salinity and the broader chemical complexity of oilfield wastewater, which generally demands more intensive pretreatment than reclaimed municipal water before it can approach the quality thresholds described in the ACWA-EPA guidance.

Regulatory and Permitting Hurdles Still Facing Operators

Federal attention to the data center-water relationship has been building for some time. A primer published by the Congressional Research Service frames data center water use and water-sourcing questions as an active policy area, reflecting how quickly this issue moved from a sustainability footnote to a subject of congressional interest. That kind of federal-level framing tends to precede, rather than follow, state and local permitting changes.

At the local level, water storage and sourcing decisions are already shaping up to be as contentious as power hookups. Zoning fights like the one chronicled in coverage of Upper Mount Bethel’s data center rules vote show how quickly municipal boards are writing new conditions into permits, and produced-water storage, with its own fire code, corrosion, and wastewater-discharge questions, is a natural candidate for the next wave of site-specific conditions.

What Industry Voices Are Saying About the Permian Basin’s Produced-Water Bet

Discussion of pairing West Texas’s oil and gas infrastructure with data center cooling needs has moved from theoretical to practical among people close to the Permian Basin’s energy economy. As Walker put it in a Houston Chronicle report on the subject, “to the extent that the data centers are going to use water regardless, I would rather them use treated produced water than other water sources.” That framing, treated produced water as a lesser-conflict alternative to drawing down local freshwater, is becoming the industry’s default argument for the practice.

Smith, quoted in the same Houston Chronicle report, pointed to the regional logic behind the pairing directly: “If you go to West Texas, you can get your power source from natural gas and your cooling source from produced water.” The comment captures why the Permian is drawing disproportionate attention from data center site selectors right now, the region offers both halves of the energy-water equation in the same geography, which is unusual outside of oil and gas country.

Market Impact: Tank Makers, Water Treatment Vendors, and Hyperscalers

For tank manufacturers and water treatment vendors, produced-water storage at data centers is a genuine new demand channel layered on top of existing oil and gas and municipal utility business. CST Industries, TanksCo, and comparable bolted-steel and glass-fused-to-steel tank builders are now positioning data center water infrastructure as a distinct product category rather than a side application, which tends to accelerate standardization and, eventually, price competition.

For hyperscalers and colocation operators, the appeal is straightforward: a water source that insulates a project from the kind of public pushback already visible in cases like the India AI hub water dispute, and from the growing scrutiny tracked in coverage of the broader AI data center power and resource crisis. Produced water will not replace municipal or reclaimed supply outright, but as one input in a blended water strategy, it gives site selectors another lever to pull when a region’s freshwater politics get difficult.

Historical Context: From Freshwater Guzzlers to Water-Smart AI Infrastructure

Data centers have spent the better part of two decades defaulting to municipal water for evaporative cooling because it was simple, cheap, and politically invisible. That calculus changed once AI training and inference workloads pushed facility power densities, and therefore cooling loads, well beyond what most regions anticipated when their water infrastructure was built. The result has been a scramble across the industry: utilities negotiating massive capital plans like the one detailed in reporting on Duke Energy’s $103 billion data center power plan, operators investing in liquid cooling to cut consumption, and now, serious consideration of nontraditional water sources that would have been dismissed a few years ago.

Produced water’s journey from disposal liability to potential cooling resource mirrors a broader pattern in data center infrastructure: constraints that once seemed fixed, land, power, now water, keep getting reengineered once the economics of AI compute make it worthwhile.

Competitive Landscape: Who’s Building the Storage Infrastructure

The competitive field for produced-water storage infrastructure splits along two lines: bolted-steel tank builders like TanksCo, who bring NFPA 22 compliance and oil-and-gas-proven designs, and glass-fused-to-steel specialists like CST Industries, who bring corrosion resistance tuned for utility-grade water storage. Neither camp has a monopoly on the data center use case yet, and expect a wave of product-line extensions, Aquastore-branded systems rated specifically for high-salinity service, or NFPA-compliant tanks marketed with cooling-grade treatment add-ons, as both sides chase the same emerging demand.

Storage & Consumption BenchmarkFigureSource
Bolted-steel tank capacity range15,000 to 1,000,000+ gallonsTanksCo
Uptime Institute onsite makeup-water reserveApproximately 12 hoursProTank / Uptime Institute Tier Standard
Tier III storage requirementConcurrently maintainable, no capacity drop below thresholdProTank / Uptime Institute Tier Standard
Typical evaporative cooling consumptionApproximately 400,000 gallons per megawatt annuallyProTank
Fire-protection standard for bolted-steel tanksNFPA 22 complianceTanksCo
Data center product linesAquastore, TecTank, OptiDomeCST Industries

Predictions: Where Produced-Water Reuse Goes From Here

  • Expect tank manufacturers to introduce produced-water-specific product lines with corrosion coatings and treatment packages bundled in, rather than leaving pretreatment entirely to third-party vendors.
  • Permian Basin data center announcements will increasingly cite produced water sourcing as a differentiator in public messaging, following the same playbook already used for onsite natural gas power.
  • Water quality standards currently written for reclaimed municipal water, including the turbidity, disinfectant, and Legionella thresholds detailed above, will likely be adapted into produced-water-specific guidance within the next development cycle rather than written from scratch.
  • Regulatory scrutiny will intensify faster than technical standards can mature, mirroring the pattern already seen in local zoning disputes over data center power and water use.
  • Hybrid water strategies, blending produced water, reclaimed municipal supply, and reduced-consumption liquid cooling, will become the default pitch for new AI data center campuses in water-stressed regions rather than any single source dominating.

Related

FAQ

What is produced water?
Produced water is high-salinity wastewater generated as a byproduct of oil and gas drilling operations. It has historically been reinjected underground or disposed of, rather than reused.

Why are data centers interested in produced water for cooling?
Produced water does not compete with municipal drinking supply or agricultural use, which makes it attractive in drought-stressed regions where freshwater allocation for new AI data centers has become politically contentious.

How much water does a data center actually use for cooling?
According to ProTank, facilities relying on evaporative cooling consume roughly 400,000 gallons per megawatt of IT capacity annually, a figure that scales quickly across hundreds of megawatts at a hyperscale campus.

What storage capacity do data centers need for backup water?
The Uptime Institute Tier Standard calls for approximately 12 hours of onsite makeup-water storage at evaporative-cooling facilities, per ProTank’s guidance, with Tier III certification requiring that storage be concurrently maintainable.

What water quality standards would produced water need to meet?
Guidance for reclaimed water reuse at data centers specifies turbidity below 3 NTU, total suspended solids under 5 mg/L, E. coli nondetectable in 75% of samples, Legionella pneumophila below 100 CFU/mL, and specific chlorine and disinfectant residual minimums in the pipeline and cooling units.

What tank types are used to store produced water at data centers?
Bolted-steel tanks built to NFPA 22 fire-protection standards, offered by manufacturers like TanksCo in sizes from 15,000 to over 1,000,000 gallons, and glass-fused-to-steel systems such as CST Industries’ Aquastore line, are both in use for this application.

Why is the Permian Basin a focus for produced-water cooling?
The region combines abundant natural gas for onsite power generation with some of the highest produced-water volumes from decades of oil drilling, giving data center developers both halves of the energy-water equation in one location.

Is produced water safe for direct use in evaporative cooling towers?
Not without treatment. Because evaporative towers aerosolize water, any source used must be treated to meet strict biological and chemical thresholds, including Legionella and E. coli limits, before it can safely enter that equipment.

Related Coverage

Marcus Chen

Marcus Chen

Gaming & Consumer Tech Editor

Marcus Chen is a senior editor at Tech Insider, where he leads coverage of the US online gaming market, including sweepstakes and social casinos, alongside consumer technology. He evaluates operators on their published terms, licensing and RNG certifications, stated redemption policies, and corroborating independent reporting, and writes plainly about what the evidence supports. Tech Insider does not run first-party money tests and does not gamble with reader funds. Marcus has reported on the technology and online-gaming industries for more than a decade.

View all articles