Why Queretaro Forces a Non-Discharge Cooling Water Strategy
A hyperscale data center in Santiago de Queretaro cannot default to a "pretreat and discharge" cooling-tower design. Eight of the state's eleven aquifers are overexploited, and the Valle de San Juan del Río aquifer recorded a 56.8 billion-litre deficit in July 2025, according to Expansion, cited in Mexico Business News (2025-11). The same reporting documents Viborillas residents receiving municipal water three days a week, with month-long outages in extreme cases, in communities immediately adjacent to new hyperscale clusters. This is not a future risk: it is the operating environment on day one of mechanical commissioning.
Queretaro hosts at least 20 operational data centres commanding 79% of Mexico's installed capacity, and the Mexican Data Centre Association (MEXDC) projects national capacity growth from 200 MW today to 1.5 GW by 2030, per Mexico Business News (2025-11). The 2022 Queretaro "Water Law" allows concession of water services to private companies, and the state government routinely exempts industrial-park data centres from standard environmental impact reports, according to the same source. Permitting therefore favours a design that demonstrates water stewardship, not a design that relies on a generous sewer or stormwater discharge consent.
The practical engineering consequence is that blowdown design flow must be sized for reuse and concentration, not for downstream sewer capacity. End-of-pipe discharge is treated as a contingency, not a baseline. AWS has signalled this by confirming air-based cooling for its Queretaro region, per Paula Bellizia, AWS, speaking to Expansion (2025-11).
What Cooling Tower Blowdown Looks Like in a Queretaro Hyperscale Plant
A typical cooling tower operating at 4 cycles of concentration loses 25–30% of makeup water to blowdown, per Genesis Water Technologies. A facility drawing 10 million US gal/month of makeup therefore generates 2.5–3 million US gal/month of blowdown. Scaling in hyperscale design typically means hundreds of thousands to millions of gallons per day of blowdown, all of which must be reused, evaporated, or hauled.
Blowdown chemistry is the constraint that drives equipment selection. Per Genesis Water Technologies, the relevant parameters are:
- TDS: 1,200–6,000 mg/L, roughly 4–8× the makeup water, depending on cycles of concentration and source water quality.
- Suspended solids: 10–50 mg/L, including corrosion products, biofilm fragments, and airborne particulates that pass basin filtration.
- Treatment chemicals: biocides, corrosion and scale inhibitors, dispersants accumulate in the blowdown stream and must be neutralised or removed before reuse or discharge.
- Scaling minerals: concentrated calcium, magnesium, silica, and alkalinity drive precipitation risk in any downstream membrane or thermal system.
Pushing cycles of concentration higher reduces blowdown volume but worsens every downstream parameter: scaling potential rises, membrane recovery drops, biocide residual increases, and the concentrate stream becomes harder to treat economically. The design target is therefore the highest cycles of concentration that the makeup chemistry and downstream treatment train can sustain together.
| Parameter | Typical Range at 4 COC | Source |
|---|---|---|
| Blowdown fraction of makeup | 25–30% | Genesis Water Technologies |
| TDS | 1,200–6,000 mg/L | Genesis Water Technologies |
| Suspended solids | 10–50 mg/L | Genesis Water Technologies |
| Cycles of concentration | 4 (typical), higher with side-stream filtration | Genesis Water Technologies |
Three Treatment Trains That Fit a Queretaro Site
For a Queretaro hyperscale site, three end-of-pipe strategies are technically credible, with a fourth upstream option that eliminates blowdown entirely. Each has a distinct cost, recovery, and permitting profile.
Option 1 — Cooling-tower makeup reuse (side-stream filtration + RO). Side-stream filtration at 1–5% of circulation flow drops suspended solids to manageable levels, after which an industrial RO skid for blowdown recovery returns permeate at 10–50 mg/L TDS to the cooling loop, achieving 60–85% overall recovery, per Genesis Water Technologies. A 50,000 GPD RO system treating blowdown carries CAPEX of US$250,000–500,000 installed and OPEX of US$1.50–3.00 per thousand gallons, covering energy, antiscalant, and membrane replacement. A multi-media filter for turbidity and TSS cut ahead of the membranes is the standard pretreatment configuration.
Option 2 — Partial ZLD (RO + MVC on the concentrate). The RO concentrate is fed to a Mechanical Vapor Compression (MVC) brine concentrator that produces distillate below 10 mg/L TDS. Per Genesis Water Technologies, this combination reaches 85–95% overall recovery, with MVC CAPEX of US$1–3 million for 10,000–30,000 GPD systems and energy use of 15–25 kWh per 1,000 US gal of distillate.
Option 3 — Full ZLD (RO + evaporator + crystallizer). A complete ZLD train adds a crystallizer to the partial-ZLD configuration, producing a solid salt cake for disposal. Per Genesis Water Technologies, full ZLD delivers 95–99% recovery, with CAPEX of US$3–8 million and OPEX of US$5–15 per thousand gallons. The economics only close where discharge is genuinely impossible and water is more expensive than energy.
Option 4 — Eliminate blowdown upstream. Air-based or closed-loop cooling, as AWS has confirmed for its Queretaro facilities, removes the blowdown stream entirely. This is a CAPEX-versus-OPEX trade against end-of-pipe ZLD: higher facility CAPEX in exchange for a discharge-free envelope. Turner & Townsend, via Mexico Business News (2025-11), notes that newer facilities adopt closed-loop and air-based systems to eliminate freshwater consumption.
Nanofiltration is an alternative to RO when hardness rather than total TDS drives the discharge or reuse constraint, operating at 75–150 psi with 70–85% recovery and permeate TDS at 30–50% of feed, per Genesis Water Technologies.
| Strategy | Overall Recovery | CAPEX Band | OPEX Band | Permit Posture |
|---|---|---|---|---|
| Side-stream filtration + RO (reuse) | 60–85% | US$250k–500k (50,000 GPD RO) | US$1.50–3.00/kgal | Reuse-first; small brine stream remains |
| Partial ZLD (RO + MVC) | 85–95% | US$1–3M (10k–30k GPD MVC) | 15–25 kWh/kgal distillate | 5–10% liquid residue; concentrate hauling possible |
| Full ZLD (RO + evaporator + crystallizer) | 95–99% | US$3–8M | US$5–15/kgal | Discharge-free; solid salt cake for disposal |
| Air-based / closed-loop cooling (upstream) | ~100% (no blowdown) | Higher facility CAPEX | Lower water OPEX, higher energy | Eliminates discharge permit need |
All cost and recovery bands are quoted from Genesis Water Technologies (2025) and are based on standard data centre blowdown applications; site-specific quotations should be requested for an EPC bid.
Sizing the Side-Stream Filtration + RO Reuse Loop
For a reuse-first design, the equipment train follows a fixed sequence. Side-stream filtration handles 1–5% of total circulation flow, removes suspended solids to 10–25 microns, and stabilises blowdown quality before the membranes, per Genesis Water Technologies. UF pretreatment ahead of the RO train at 0.01–0.1 micron pore size handles biological and colloidal loading, with 90–95% recovery and permeate backwash. RO operates at 150–400 psi to overcome the osmotic pressure of the concentrate, with PLC-controlled antiscalant and biocide dosing to prevent silica and calcium scaling.
The pretreatment targets are non-negotiable. Feed water Silt Density Index (SDI) should be held below 3, pH adjusted for membrane chemistry, and turbidity cut to below 1 NTU. A multi-media filter upstream is the most common configuration to meet these targets, per Genesis Water Technologies. Without them, RO membrane life collapses and recovery rates fall short of the 60–85% design point.
The reuse loop returns permeate to the cooling tower as high-quality makeup, which raises the blended cycles of concentration and reduces total makeup draw from the aquifer. The concentrate stream — typically 15–40% of the feed — is the only liquid that still requires disposal or further treatment.
NOM-001 and the Queretaro Discharge Reality
Federal industrial discharge in Mexico is governed by NOM-001-SEMARNAT-2022, which sets parameter-specific limits for heavy metals, TSS, BOD, oils and greases, and total dissolved solids depending on the receiving body. Per Genesis Water Technologies, some jurisdictions have implemented TDS caps below 1,500 mg/L, which effectively prohibits discharge of concentrated blowdown without treatment.
In Queretaro, even where NOM-001 limits are technically met, the political and social envelope makes routine discharge a reputational risk. Viborillas residents are organising through groups such as Voceras de la Madre Tierra, and a MORENA Deputy has proposed a special commission to investigate private water concessions, per Mexico Business News (2025-11). The 2022 Queretaro Water Law allows concession of water services to private companies, and the state government routinely exempts industrial-park data centres from standard environmental impact reports, per the same source. The structural conclusion is that the permit pathway favours designs that minimise or eliminate discharge, regardless of whether NOM-001 limits are met.
Any blowdown design should therefore be evaluated on a reuse-first basis, with discharge sized for emergency bypass and startup only. Before sizing any discharge pipework, request a current site-specific concession title and CONAGUA discharge permit from the EPC; both have been delayed or withheld for recent Queretaro projects, and proceeding without them exposes the developer to enforcement risk.
Decision Framework: Reuse, Partial ZLD, or Full ZLD for Queretaro
The choice between strategies should be driven by site-specific water risk, not by vendor preference. The decision rules below assume that the site has a stable aquifer concession and that the developer is willing to accept a smaller residual liquid stream where unavoidable.
Choose reuse (side-stream UF + RO) when: the site has a stable aquifer concession, the developer will accept discharge of an 80–85% smaller brine stream, and the project budget favours lower CAPEX and OPEX over a discharge-free envelope. This is the lowest-cost option of the three end-of-pipe strategies, per Genesis Water Technologies.
Choose partial ZLD (RO + MVC on concentrate) when: concentrate hauling or deep-well injection is restricted, but the site can absorb 5–10% residual liquid waste and has reliable power. This is the middle-ground option, per Genesis Water Technologies, and is well suited to sites near the Viborillas–El Marques–Colon corridor where community scrutiny is highest.
Choose full ZLD when: the receiving aquifer is in a CONAGUA deficit zone (such as the Valle de San Juan del Río), the project faces community opposition comparable to Viborillas, or the developer wants a single permit envelope that is genuinely discharge-free. Per Genesis Water Technologies, full ZLD CAPEX runs US$3–8M and OPEX US$5–15/kgal, and the economics only close where water is more expensive than energy.
Choose air-based or closed-loop cooling upstream — as AWS has confirmed for Queretaro — when the project budget supports higher facility CAPEX in exchange for eliminating the blowdown stream entirely, per AWS via Expansion (2025-11). For a deeper review of ZLD economics and design choices, see the ZLD plant design and economics guide. For a regional comparison on a different stressed aquifer, see the Latin American data center blowdown case study on Valencia, Venezuela.
Frequently Asked Questions
What CAPEX should we budget for a side-stream filtration + RO reuse train on a Queretaro hyperscale site?
For a 50,000 GPD RO skid treating cooling-tower blowdown, CAPEX is US$250,000–500,000 installed, with OPEX of US$1.50–3.00 per thousand gallons covering energy, antiscalant, and membrane replacement, per Genesis Water Technologies. Add the side-stream filtration and UF pretreatment to scope the full train, and request a site-specific quotation tied to your blowdown flow, TDS, and cycles-of-concentration target before locking budget.
How do we choose between reuse, partial ZLD, and full ZLD without overpaying for capacity we don't need?
Match the strategy to site water risk rather than to vendor preference. Reuse via side-stream UF + RO is the lowest-cost option where a small residual brine stream is acceptable. Partial ZLD (RO + MVC) is the middle ground where concentrate hauling is restricted. Full ZLD is justified only where the receiving aquifer is in CONAGUA deficit, community opposition is acute, or a discharge-free permit envelope is a hard requirement, per Genesis Water Technologies. For each tier, request a recovery-rate and CAPEX/OPEX band tied to your actual blowdown flow rather than generic curves.
What compliance inputs do we need from the EPC and CONAGUA before sizing discharge pipework?
Request a current site-specific concession title and a CONAGUA discharge permit from the EPC before any pipe sizing. Both have been delayed or withheld for recent Queretaro projects, and the state government routinely exempts industrial-park data centres from standard environmental impact reports, per Mexico Business News (2025-11). Without these documents, a discharge design is a paper exercise; sizing should be treated as emergency-bypass capacity only.
What supplier and lead-time risks should we flag for the RO and MVC equipment on a 2026 build?
Lead time for an industrial RO skid at hyperscale flows is dominated by membrane delivery, and MVC units are typically built to order; request a current factory lead time and a membrane-replacement schedule from the supplier before signing the EPC contract. For Mexican sites, confirm the supplier has a local service partner with stocked spares and authorised start-up engineers, and ask for a reference installation in Mexico or Latin America. Quote both CAPEX and OPEX bands in the bid (US$1.50–3.00/kgal for RO and 15–25 kWh/kgal distillate for MVC, per Genesis Water Technologies) so that operating-cost risk stays with the supplier if actual performance falls short.
