Why Cooling-Tower Blowdown, Not Sanitary Sewage, Drives Tijuana Data-Center Compliance
Cooling-tower blowdown is the stream that carries the regulatory and operational load at a Tijuana data center, not sanitary sewage, condensate, or office greywater. It is the bleed-off that protects the recirculating cooling loop from excessive accumulation of dissolved solids, and it concentrates everything the loop has been exposed to: scaling minerals (calcium, magnesium, silica, alkalinity), treatment chemicals (biocides, corrosion inhibitors, scale inhibitors, dispersants), suspended solids, and corrosion byproducts. Genesis Water Tech (2026) characterizes blowdown as having total dissolved solids (TDS) 4–8 times higher than makeup water — typically 1,200 to 6,000 mg/L — plus suspended solids of 10–50 mg/L and biological content that has to be addressed in any recovery or discharge system.
A cooling tower operating at 4 cycles of concentration loses approximately 25–30% of makeup water as blowdown; for a facility using 10 million gallons monthly this is 2.5–3 million gallons discharged (Genesis Water Tech, 2026). Because the blowdown stream is concentrated, small volumetric mistakes in design create large contaminant loads downstream — and that is what regulators care about.
Under Mexican federal rules, any industrial discharge to a sewer or surface water body must meet NOM-001-SEMARNAT (current 2021 text), and the local operator — in Tijuana, the Comisión Estatal del Agua (CEA) and the municipal utility CESPT — can layer additional pretreatment limits on top. A second, often overlooked compliance stream is glycol from closed loops (outdoor piping, dry coolers, some liquid-cooling loops). Both ethylene and propylene glycol create high oxygen demand in surface water if released, lowering dissolved oxygen and harming aquatic life, so containment and documented disposal through a licensed ITAD or hazardous-waste partner are part of the compliance picture (HOBI, 2026).
Before sizing anything, confirm where each stream actually goes on this site: blowdown to sewer or to a reuse skid, condensate to landscape or drain, glycol to licensed disposal, and sanitary to the CESPT sanitary line. Conflating those streams is the most common source of permitting delay on Mexican data-center projects.
Tijuana-Specific Constraints: Water Stress, NOM-001, and the Rio Colorado Allocation
Tijuana draws a significant share of its municipal supply from the Río Colorado allocation through the Mesa de Otay system, and that system is chronically over-allocated. For a 2026 data-center build, that means two things: first, the engineer must design against a stressed source and any drought decree in force at the time of procurement, not against a hypothetical nominal supply; second, the case for reuse or ZLD often stands on water stewardship and on the client's published ESG commitments, not only on the law.
Federal surface-water and sewer discharge is governed by NOM-001-SEMARNAT-2021, which sets pollutant limits by receiving-body type. The current text is the operative reference, but specific numerical limits and any 2026 updates must be verified with a local environmental consultant before specification — they are not reproduced in this guide. CESPT applies additional sewer-pretreatment limits — analogous to the published Jacksonville Electric Authority example cited in HOBI (2026) for cooling-tower blowdown — covering pH, heavy metals, residual chlorine, oil and grease, and frequently TDS. The exact local numerical limits must be requested from CESPT in writing before any train is sized.
Most new Tijuana builds target cloud or AI workloads, so even when Mexican law would allow sewer discharge, hyperscaler ESG mandates often force reuse or ZLD regardless of regulation (HOBI, 2026). That tension — looser local pretreatment versus stricter client ESG — is the single biggest variable that separates a Tijuana design from a US design, and it is missing from most generic guides.
Inputs the engineer must gather before sizing: the CNA title concession volume, the current CEA allocation, CESPT pretreatment limits in writing, the data-center client's published water and ESG commitments, and any drought-decree operating restrictions in force at the expected commissioning date.
The Three End-Use Strategies: Reuse, Discharge Compliance, and ZLD
The strategic decision — reuse, discharge compliance, or zero liquid discharge (ZLD) — drives everything downstream, including CAPEX, OPEX, and which block diagram the EPC eventually issues for construction. Genesis Water Tech (2026) frames the three end-uses as fundamentally different economic stories, not as points on a continuum.
Cooling-tower makeup reuse typically achieves 60–85% recovery and is the most common objective for new data centers; the permeate blends back into the cooling system, reducing both makeup demand and discharge volume. Discharge compliance is justified by avoided discharge fees rather than by water savings — Genesis Water Tech (2026) notes that direct discharge fees in water-stressed regions already exceed $5–15 per 1,000 gallons — and in Tijuana the avoided fee stack includes CESPT volumetric charges plus any NOM-001 monitoring cost. ZLD achieves 95–99% overall recovery by combining RO (70–80% recovery) with mechanical vapor compression (MVC, 95% recovery of concentrate) and a crystallizer, with data-center-scale CAPEX of $3–8 million and OPEX of $5–15 per 1,000 gallons treated (Genesis Water Tech, 2026). A middle path is partial ZLD: concentrate blowdown to reduce volume by 80–90%, then deep-well inject, haul, or discharge the residual under a special permit (Genesis Water Tech, 2026).
For a Tijuana facility, the choice is rarely purely economic: ESG mandates from cloud clients and the local water-stress narrative push toward reuse or ZLD, while CAPEX pressure pushes toward discharge compliance with side-stream filtration only. The decision rule used in the field: if the client has a published water-positive or WUE target, plan for reuse plus partial ZLD; if the client is a domestic colocation operator, plan for side-stream filtration and discharge compliance against CESPT limits.
| Strategy | Typical recovery | Primary value driver | Best Tijuana fit |
|---|---|---|---|
| Cooling-tower makeup reuse | 60–85% | Reduced makeup + reduced discharge | Hyperscaler builds with published WUE target |
| Discharge compliance | N/A (treatment to permit) | Avoided CESPT and NOM-001 fees | Domestic colocation, lower ESG pressure |
| Full ZLD | 95–99% | No liquid waste stream, ESG narrative | Contractual ZLD requirement or discharge ban |
| Partial ZLD | 80–90% volume reduction | Lower CAPEX than full ZLD, lower disposal cost | Middle path when reuse is needed but full ZLD uneconomic |
Matching Cooling Design to Treatment Train
The cooling system the facility is actually buying determines the blowdown stream the treatment train has to handle, and that mapping is missing from most generic guides. Each design generates a different stream with a different contaminant load and a different peak-to-average ratio.
A water-cooled facility with an evaporative cooling tower at 4 cycles of concentration is the reference case: blowdown typically carries TDS 1,200–6,000 mg/L, suspended solids 10–50 mg/L, and the full cooling-water chemistry package (Genesis Water Tech, 2026). A hybrid wet/dry system reduces blowdown volume proportionally to its dry-mode duty cycle, which is useful when the client wants reuse but not full ZLD — the train must be sized for peak wet-mode flow, not average. A closed-loop glycol system does not generate blowdown during normal operation, but spent glycol from drains, maintenance, or accidental release must be handled by a licensed ITAD or hazardous-waste partner (HOBI, 2026). An adiabatic or immersion-cooled facility still produces a smaller blowdown stream from the secondary loop plus any makeup-side concentration; pretreatment can be lighter but the chemistry still concentrates, so a smaller RO skid plus UV polishing is often the right fit.
Inputs the engineer must request before specifying the train: cooling system type and topology, design cycles of concentration, expected annual blowdown volume in L/s or m³/h, peak wet-mode flow, and the full list of treatment chemicals used in the cooling program — including biocide, scale inhibitor, and corrosion inhibitor chemistry, because that chemistry drives both membrane compatibility and discharge permit limits.
| Cooling design | Blowdown character | Treatment implication |
|---|---|---|
| Evaporative tower, 4 COC (reference) | TDS 1,200–6,000 mg/L, SS 10–50 mg/L, biocides, scale inhibitors | Full side-stream filtration → UF → RO train sized for full blowdown |
| Hybrid wet/dry | Variable; depends on dry-mode duty cycle | Size for peak wet-mode flow, not average |
| Closed-loop glycol | No routine blowdown; spent glycol on maintenance | Containment + licensed ITAD disposal; no continuous train |
| Adiabatic / immersion | Smaller secondary-loop blowdown, chemistry still concentrated | Lighter pretreatment; smaller RO skid + UV polish |
Core Treatment Technologies and How They Combine
Treat the train as a connected chain of blocks with defined roles, not as a vendor menu. Each block has a known CAPEX/OPEX band from the Genesis Water Tech (2026) dataset and a defined place in the sequence.
Side-stream filtration at 1–5% of circulation flow is the first block. Self-cleaning spiral or multimedia filters to 10–25 microns reduce suspended solids and biological loading before the water reaches the membranes; CAPEX is $50,000–$200,000 for typical data-center flow (Genesis Water Tech, 2026). Ultrafiltration (0.01–0.1 micron PVDF) follows, removing suspended solids, bacteria, and high-molecular-weight organics at 10–30 psi with 90–95% recovery, and serves as the standard RO pretreatment (Genesis Water Tech, 2026) — a hollow-fiber UF system in this position. Reverse osmosis removes 95–99% of TDS, hardness, and silica at 150–400 psi, with 50–85% recovery limited by scaling; permeate at 10–50 mg/L TDS is suitable for direct cooling-tower makeup (Genesis Water Tech, 2026). An industrial RO skid at 50,000 GPD installed costs $250,000–$500,000 with OPEX of $1.50–$3.00 per 1,000 gallons. A multi-media filter upstream of the UF keeps the membrane feed within the 10–15 micron target and extends backwash intervals.
Nanofiltration (75–150 psi, 70–85% recovery) is the right choice when hardness and sulfate drive the discharge problem but full TDS removal is not needed; permeate TDS is typically 30–50% of feed (Genesis Water Tech, 2026). Mechanical vapor compression produces 95–98% recovery and less than 10 mg/L TDS distillate, with CAPEX of $1–3 million for 10,000–30,000 GPD and energy consumption of 15–25 kWh per 1,000 US gallons — only viable when waste heat is unavailable or when ZLD is contractually required (Genesis Water Tech, 2026).
A common train for a 2026 Tijuana hyperscale site is side-stream filtration → UF → antiscalant-conditioned RO → permeate to cooling-tower makeup, with the RO concentrate routed to MVC only if ZLD is contractually required. A lamella clarifier or DAF upstream of UF is sometimes justified when corrosion byproducts or biological floc are heavy; design at 4–300 m³/h depending on sidestream flow.
Sizing a Tijuana Reference Case: 5 MGD Makeup, 4 Cycles of Concentration
At 5 million gallons per day of cooling-tower makeup and 4 cycles of concentration, blowdown is approximately 25–30% of makeup — i.e., 1.25–1.5 MGD (Genesis Water Tech, 2026). The exact percentage depends on drift and windage losses, which the cooling-tower vendor should supply.
The first lever is operational, not capital: raising cycles of concentration from 3 to 6 cuts makeup by about 20% and blowdown by about 50%, per DOE guidance cited in HOBI (2026) — a higher-leverage operating change than adding a new membrane skid. Antiscalant injection and a side-stream filtration skid are the cheapest first dollars; they enable higher cycles of concentration and reduce fouling on the RO membranes.
A 50,000 GPD industrial RO skid at $250,000–$500,000 installed handles only the reusable fraction. A 1.25 MGD blowdown stream with 75% reuse needs roughly 19 RO skids of that size, so a multi-skid or larger single RO train is the realistic scale, with CAPEX scaling in proportion to the Genesis Water Tech (2026) dataset range. A full ZLD package at $3–8 million (Genesis Water Tech, 2026) becomes worth modelling when the client requires 95–99% recovery or when CESPT pretreatment limits effectively prohibit any discharge — request the full CAPEX stack (RO train + MVC + crystallizer + brine handling) and the OPEX of $5–15 per 1,000 gallons before committing.
Use the Open Engineering / EngrXiv 2026 LCA finding: a reclaimed-water UF+RO train has about five times the energy demand of freshwater, but the indirect water penalty is only about 0.93 L/m³ — the trade-off is energy and GWP, not water security (Cartagena Vaca et al., 2026).
Capital, Operating, and Environmental Trade-Offs in 2026
The CAPEX/OPEX bands from Genesis Water Tech (2026) are the working numbers for a 2026 budget. Side-stream filtration runs $50,000–$200,000; an RO skid at 50,000 GPD is $250,000–$500,000; MVC for 10,000–30,000 GPD is $1–3 million; and a full data-center-scale ZLD package is $3–8 million. OPEX bands are: RO at $1.50–$3.00 per 1,000 gallons, ZLD at $5–15 per 1,000 gallons, and direct discharge fees in water-stressed regions at $5–$15 per 1,000 gallons — meaning ZLD OPEX and avoided discharge fees are often within the same order of magnitude.
The environmental trade-off from the EngrXiv 2026 LCA (Cartagena Vaca et al., 2026) is the headline number for the sustainability narrative: a reclaimed-water train has about twice the GWP of freshwater, and a UF+RO scenario drives energy demand to over five times that of freshwater, but the indirect water footprint is under 0.1% of the freshwater displacement benefit. In a fully decarbonized grid scenario, the GWP penalty of reuse becomes negligible while the water savings continue to accrue at full value through the asset's service life (Cartagena Vaca et al., 2026).
A chemical dosing skid is required for antiscalant, biostat control, and CIP; an automatic antiscalant dosing skid sized to the RO train and integrated with the cooling-water chemistry program prevents chemical accumulation in blowdown. A sludge-handling line — for example, a plate-and-frame filter press for RO CIP waste and clarifier underflow — is usually small in flow but should be in scope from day one to avoid permit complications later.
| Block | CAPEX (data-center scale) | OPEX | Source |
|---|---|---|---|
| Side-stream filtration | $50,000–$200,000 | Solids disposal + maintenance | Genesis Water Tech, 2026 |
| RO skid (50,000 GPD) | $250,000–$500,000 | $1.50–$3.00 / 1,000 gal | Genesis Water Tech, 2026 |
| MVC (10,000–30,000 GPD) | $1–3 million | 15–25 kWh / 1,000 gal distillate | Genesis Water Tech, 2026 |
| Full ZLD package | $3–8 million | $5–15 / 1,000 gal treated | Genesis Water Tech, 2026 |
| Direct discharge fee (water-stressed) | N/A | $5–$15 / 1,000 gal | Genesis Water Tech, 2026 |
Frequently Asked Questions
What CAPEX and OPEX should a Tijuana data center budget for cooling-tower blowdown treatment in 2026?
Budget on the bands from the Genesis Water Tech (2026) dataset: $50,000–$200,000 for side-stream filtration, $250,000–$500,000 for a 50,000 GPD RO skid, $1–3 million for 10,000–30,000 GPD MVC, and $3–8 million for a full data-center ZLD package. OPEX runs $1.50–$3.00 per 1,000 gallons for RO and $5–15 per 1,000 gallons for ZLD, against avoided discharge fees of $5–$15 per 1,000 gallons in water-stressed regions. For a Tijuana site, scale these to your actual 1.25–1.5 MGD blowdown at 5 MGD makeup and 4 COC, and request vendor quotations tied to feed-water chemistry rather than book ranges.
How do I select between reuse, discharge compliance, and ZLD for a specific Tijuana site?
Ask four questions before selecting: (1) what is the client's published water or WUE target, if any; (2) what is the CNA title concession volume versus projected demand, and is a drought decree in force; (3) what are the CESPT pretreatment limits in writing for TDS, heavy metals, residual chlorine, and oil and grease; and (4) is there hyperscaler ESG language that contractually forces reuse or ZLD. If the client has a WUE or water-positive target, plan reuse plus partial ZLD; if the client is a domestic colocation operator, plan side-stream filtration and discharge compliance against CESPT limits. The geographic and ESG inputs are qualitative and must be confirmed in writing before procurement — no vendor should size a train without them.
Does a closed-loop or glycol system still need blowdown treatment?
No continuous blowdown treatment is required for a closed-loop glycol system during normal operation, but the site still needs a containment and disposal plan. Both ethylene and propylene glycol create high oxygen demand in surface water and can lower dissolved oxygen if released; ethylene glycol also carries higher toxicity concerns (HOBI, 2026). Spent glycol from drains, maintenance, or accidental release must be handled by a licensed ITAD or hazardous-waste partner for documented, compliant removal.
What Mexican permits and limits apply to cooling-tower blowdown in Tijuana?
Any industrial discharge to a sewer or surface water body must meet NOM-001-SEMARNAT-2021 (current text), and CESPT applies additional sewer-pretreatment limits analogous to the published Jacksonville Electric Authority example cited in HOBI (2026) for cooling-tower blowdown. Specific numerical limits are not reproduced in this guide and must be requested in writing from CESPT and verified with a local environmental consultant before specification. The engineer also needs the CNA title concession volume, the current CEA allocation, and any drought-decree operating restrictions in force at the expected commissioning date.
How does a UF+RO blowdown train compare with a freshwater baseline on energy and water footprint?
Per the EngrXiv 2026 LCA (Cartagena Vaca et al., 2026), a reclaimed-water train has about twice the GWP of the freshwater case, and a UF+RO scenario drives energy demand to over five times that of freshwater — but the indirect water penalty is only about 0.93 L/m³, or less than 0.1% of the direct displacement benefit. Under a fully decarbonized grid scenario, the GWP penalty becomes negligible while water savings continue to accrue at full value through the asset's service life. The trade-off is energy and carbon, not water security.
Related Equipment
- industrial RO skid — specifications, capacity range, and technical data
