Why Tashkent Is a Stand-Alone Design Problem in 2026
A 2026 Tashkent data center must be sized to the Chirchiq–Bozsu source envelope (TDS 350–700 mg/L, hardness 180–300 mg/L as CaCO₃, silica 8–20 mg/L) and an arid-continental climate with a +38 to +40 °C summer wet-bulb and a −10 to −15 °C winter design dry-bulb, a 50 °C annual swing that a Baku, Astana, or tropical template will not resolve (per HydropureWater 2025 field benchmarks for the Chirchiq catchment and 2025 climatic reference data). The Chirchiq River and the Bozsu canal together supply Tashkent's municipal system; the same envelope feeds any greenfield data center in the Tashkent agglomeration, so the design baseline does not move between sites in the metro area. Spring snowmelt (April–May) lifts raw-water TSS to 25–60 mg/L, and a side-stream filter on 1–5% of tower circulation must drop SDI below 3 to protect downstream RO.
The climate envelope is the second forcing function. A +38 to +40 °C summer wet-bulb is closer to Baku's 26–28 °C envelope than to Astana's 18–20 °C, so evaporative loss at the tower runs at the Baku benchmark, not the 15–30% Astana reduction. The winter dry-bulb of −10 to −15 °C is milder than Astana's −30 to −35 °C, so freeze-protection logic is less aggressive, but the 50 °C annual swing still forces hybrid adiabatic / dry cooling for part of the year. Pure wet cooling in July carries a 30–40% blowdown penalty versus a hybrid reference because the dry mode carries most of the load below ~10 °C ambient (per HydropureWater Baku 2026 hybrid-cooling logic, applied to Tashkent's climate envelope).
The permit frame is dual and parallel. The State Committee for Ecology and Environmental Protection (operating under the UzNature brand for EIA scoping) issues the environmental impact assessment, and the Goswatercadastre / Basin Water Management Authority under the Ministry of Water Resources issues the water-use and discharge consent — two separate filings, both required before commissioning (per Uzbekistan environmental code structure 2025). Uzbekistan is not yet basin-classified as water-stressed at the federal level, but Tashkent sits in a transboundary Chirchiq watershed with declining per-capita renewable resource, so sovereign-cloud procurement already writes water-positive language into contracts. An MCP-01 / SanPiD-equivalent reuse standard applies when treated effluent reuses on site.
The Four Streams Every Tashkent Data Center Has to Treat
Stream separation is mechanical, not procedural: a single combined clarifier cannot hold the biology of sanitary sewage and the mineral chemistry of cooling-tower blowdown at the same time, so each stream runs on its own train and they meet only at the discharge or reuse point. The four streams and their 2026 parameter envelopes are summarised below (per HydropureWater 2025 Chirchiq field benchmarks and HydropureWater Astana 2026 four-stream framework).
| Stream | Volume / Load Envelope | Key Parameters | Treatment / Reuse |
|---|---|---|---|
| 1 — Makeup water (Chirchiq–Bozsu) | Site-total; sized to IT load + 15–30% unmeasured-loss uplift | TDS 350–700 mg/L; hardness 180–300 mg/L as CaCO₃; silica 8–20 mg/L; spring snowmelt TSS 25–60 mg/L | Side-stream side-stream multi-media filter to SDI <3 to protect RO; cooling-tower makeup |
| 2 — Cooling-tower blowdown (CTBD) | 1,800–2,100 m³/month at 5 MW and 4 CoC; 3,800–4,200 m³/month at 10 MW | TDS 2,400–6,400 mg/L at 4–6 CoC; warm 25–32 °C in summer; phosphonates, dispersants, biocides | DAF + softener/anti-scalant + BWRO; reuse as cooling-tower makeup; UzNature discharge consent for concentrate |
| 3 — Sanitary sewage | ~100 L per employee per day; pathogen-bearing | BOD₅ 200–300 mg/L; TSS 200–250 mg/L; fecal coliform present | On-site A/O package plant; fecal coliform ≤1,000 CFU/100 mL for irrigation reuse |
| 4 — Boiler blowdown / humidification bleed | Steam humidification, boiler skids; low-flow intermittent | High-TDS, silica-bearing when present | Blend into cooling-tower makeup if chemistry allows; otherwise route to CTBD train |
Stream 1 is the intake the cooling tower and any RO feed drink. It is turbid during the April–May snowmelt window and baseline hard and silica-bearing the rest of the year. A side-stream filter on 1–5% of tower circulation flow drops TSS to membrane-manageable levels and protects the downstream train from chronic fouling (per Genesis Water Tech 2025 field data). Stream 2 — cooling-tower blowdown — is the largest treatable volume on site: warm, mineralised, biocide-bearing, and the only stream with a real reuse revenue case. Stream 3 — sanitary sewage — is the smallest by volume but the most operationally sensitive because it runs every day regardless of IT load, and it must run on a dedicated anoxic/oxic (A/O) train. Stream 4 — boiler blowdown and humidification bleed — is low-flow and high-TDS; it routes to the cooling-tower makeup if the silica envelope allows, otherwise to the CTBD train.
Sizing the Cooling-Tower Blowdown Train in Tashkent's Climate
Start with the IT load and walk forward. A 5 MW facility at 4 cycles of concentration loses roughly 4–5% of circulation flow to evaporation plus 25% of makeup to blowdown; the blowdown fraction is the part that determines the size of the downstream train. At 6 CoC, blowdown drops to 1/(6−1) = 20% of makeup (per Genesis Water Tech 2025; confirmed in HydropureWater Baku 2026). The 4→6 CoC step looks like a 50% improvement on paper, but the actual blowdown-volume reduction is 5 percentage points, or about 20% — and above 5–6 CoC, microbiologically influenced corrosion, Legionella, and biofilm fouling accelerate non-linearly, so 4–6 CoC is the realistic design ceiling for an unsoftened Chirchiq feed (per HydropureWater Baku 2026, Genesis Water Tech 2025).
The climate correction cuts both ways. Tashkent's +38 to +40 °C summer wet-bulb is closer to Baku's 26–28 °C envelope than to Astana's 18–20 °C, so evaporative loss runs at the Baku benchmark rather than the 15–30% Astana reduction — but unmeasured losses (leaks, drift, emergency dumps) still add 15–30% on top of theoretical blowdown (per Genesis Water Tech 2025). Apply a 1.3–1.5 peak factor for the July–August hot window and the unmeasured-loss uplift together, and the CTBD figures land at 1,800–2,100 m³/month for a 5 MW Tashkent site and 3,800–4,200 m³/month for a 10 MW build (per HydropureWater Baku 2026 field benchmarks, adjusted for Tashkent wet-bulb). On a daily basis, a 10 MW campus should plan for 60–140 m³/day of average CTBD with peak-day allowance to 200 m³/day.
| Sizing Parameter | 5 MW Tashkent | 10 MW Tashkent | Source |
|---|---|---|---|
| IT load assumption | 5 MW | 10 MW | Per RFQ envelope |
| Cycles of concentration (CoC) | 4–6 | 4–6 | Genesis Water Tech 2025 |
| Cooling-tower makeup (monthly) | 9,000–10,500 m³/month | 19,000–21,000 m³/month | Includes 15–30% unmeasured-loss uplift |
| CTBD at 4–6 CoC (monthly) | 1,800–2,100 m³/month | 3,800–4,200 m³/month | HydropureWater Baku 2026, adjusted for Tashkent |
| Average daily CTBD | 60–70 m³/day | 130–140 m³/day | Derived |
| Peak-day CTBD (1.3–1.5×) | 80–105 m³/day | 170–200 m³/day | Peak factor 1.3–1.5 in July–August |
Winter changes the sizing logic. From December through February, a wet cooling tower in Tashkent either runs as a hybrid adiabatic / dry unit with intermittent water spray on the warmest hours, or it switches to a full air-side condenser and the cooling-tower loop is idled. Hybrid adiabatic / dry cooling can cut summer CTBD by 30–40% versus a pure-wet reference, because the dry mode carries most of the load below ~10 °C ambient (per HydropureWater Baku 2026 hybrid-cooling logic, applied to Tashkent's milder winter envelope). For a 5–10 MW Tashkent build, the defensible 2026 sizing argument is: design the wet-train equipment for the summer peak, but show the hybrid-cooling operating envelope in the UzNature submission so the regulator sees the realistic annual CTBD profile and not the worst-case summer value.
Four Treatment Trains That Cover the Tashkent Envelope
The choice of train sets the WUE the operator can report, the regulatory exposure, and the reuse revenue stream the finance team can model. Four options cover the realistic envelope for a Tashkent data center in 2026.
Option A — DAF + PLC chemical dosing. A DAF clarifier for the cooling-tower blowdown train plus a PLC-controlled chemical dosing skid lands at $35K–$60K installed. It clears UzNature discharge limits, extracts no reuse value, and has the lowest CAPEX of the four. The upgrade path to Option C is straightforward once the Goswatercadastre consent is in hand.
Option B — DAF + softener/anti-scalant + BWRO at 75–80% recovery. A twin-tank industrial water softener for CoC lift plus a BWRO unit on the blowdown train lands at $220K–$480K installed for 5–10 MW. Permeate silica ~1 mg/L is suitable for cooling-tower makeup; the concentrate is the disposal stream and needs an UzNature-acceptable sludge management plan (per Genesis Water Tech 2025 BWRO economics).
Option C — Side-stream softener + lamella clarifier + ClO₂ disinfection + partial RO on blowdown at 50–85% recovery. A lamella clarifier for footprint reduction plus a chlorine dioxide generator for cooling-loop biocide control provides 0.5–1.0 mg/L residual, preferred over chlorine because it does not form trihalomethanes and remains effective against Legionella at the 25–32 °C warm-water range. Permeate TDS 10–50 mg/L delivers 30–50% raw-water cut without full ZLD.
Option D — RO preconcentrator + MVC + crystalliser for full ZLD. $1.2M–$3.5M; reserve for sovereign-cloud builds where municipal discharge is prohibited or water-positive language is contractually binding. MVC thermal stage runs at 15–25 kWh per 1,000 US gallons with distillate below 10 mg/L TDS (per HydropureWater Baku 2026 ZLD bench; Genesis Water Tech 2025).
| Train | Core Unit Operations | CAPEX (5–10 MW) | OPEX Drivers | Reuse / Permit Posture |
|---|---|---|---|---|
| A — DAF + dosing only | DAF + PLC-controlled chemical dosing skid | $35K–$60K | Chemical dosing (~20% of OPEX); sludge haul-out | Discharge compliance only; no reuse revenue |
| B — DAF + softener + BWRO | DAF + softener + BWRO at 75–80% recovery | $220K–$480K | RO membrane replacement 3–5 yr; antiscalant; sludge haul-out | Cooling-tower makeup; Goswatercadastre consent for concentrate |
| C — Softener + lamella + ClO₂ + partial RO | Side-stream softener + lamella + ClO₂ + partial RO at 50–85% recovery | Mid-band, above B | NaCl regeneration; ClO₂ precursor; partial RO membrane cycle | Cooling-tower makeup + irrigation reuse; moderate–high permit exposure |
| D — RO preconcentrator + MVC + crystalliser (ZLD) | RO preconcentrator + MVC + crystalliser | $1.2M–$3.5M | MVC energy 15–25 kWh/1,000 gal; crystalliser media | Zero liquid discharge; closed loop; water-positive ESG |
For a 5–10 MW Tashkent greenfield in 2026, the defensible default is Option B for the first build (UzNature EIA clearance plus cooling-tower makeup reuse), with Option C sized into the civil works so the side-stream softener, lamella, and ClO₂ skids can be added at phase-2 expansion without a re-permit cycle. Option D is reserved for sovereign-cloud builds where water-positive language is written into the procurement contract. For a deeper bench on DAF design parameters under the Chirchiq envelope, see the DAF design parameters guide; for thermal brine concentration, see the MVR evaporation engineering guide.
Uzbekistan Permit Frame and Sludge Routing
Two agencies, two filings, parallel not serial. The UzNature EIA scoping and submission typically takes 60–120 days for a 5+ MW facility, anchored to the hydraulic calculations, the chemical dosing skid specification, and a sludge disposal plan (per HydropureWater Astana 2026 MENR analog, applied to Uzbekistan's environmental code structure 2025). The two most common first-submission rejection causes — missing dosing specs and missing sludge routing — drive roughly half of failed filings in analogous regimes, so the PLC-controlled chemical dosing skid selection and the dosing sequence must be locked at the UzNature design-submission step, not at final inspection.
Sludge routing must be locked at the same step: a plate-and-frame filter press for sludge dewatering, the licensed-hauler contract, and the receiving-facility permit all need to be in the UzNature file. Sludge disposal at $0.30–$0.80 per kg dry solids is the operating cost the regulator wants to see priced. The Goswatercadastre / Basin Water Management Authority under the Ministry of Water Resources issues the water-use and discharge consent once the UzNature EIA is cleared. Submittal must include hydraulic calculations, chemical dosing specifications, the receiving-waterbody or sewer-shed capacity assessment, and the sludge disposal plan. If treated effluent reuses on-site for landscaping, dust suppression, or construction water, SanPiD-equivalent reuse criteria apply — fecal coliform ≤1,000 CFU/100 mL, BOD ≤50 mg/L, turbidity <1 NTU (per HydropureWater Baku 2026 reuse standard, applied to Tashkent).
Total permit timeline 6–10 months across both submissions. Running them in parallel rather than serially is the single biggest schedule lever on a 5–10 MW greenfield. Permit costs add $4K–$10K one-time across the UzNature EIA and the Goswatercadastre consent.
CAPEX, OPEX, and Payback for a 5–10 MW Tashkent Build
Total installed CAPEX runs $35K–$60K for Option A, $220K–$480K for Option B, and $1.2M–$3.5M for Option D ZLD, with a Central-Asia logistics premium of 10–20% over the Baku reference frame applied to Uzbekistan (per HydropureWater Astana 2026). A 50,000 GPD RO unit alone lands in the $250K–$500K band installed (per Genesis Water Tech 2025). OPEX is dominated by chemical dosing (about 20% of OPEX), sludge disposal at $0.30–$0.80 per kg dry solids to a licensed hauler, and RO membrane replacement every 3–5 years (per HydropureWater Baku 2026).
| Cost Component | Option A | Option B | Option C | Option D (ZLD) |
|---|---|---|---|---|
| CAPEX band (5–10 MW installed) | $35K–$60K | $220K–$480K | Mid-band, above B | $1.2M–$3.5M |
| Central-Asia logistics uplift | +10–20% | +10–20% | +10–20% | +10–20% |
| OPEX drivers | Chemical dosing; sludge haul-out | RO membrane 3–5 yr; antiscalant; sludge haul-out | NaCl regeneration; ClO₂ precursor; partial RO membrane cycle | MVC energy 15–25 kWh/1,000 gal; crystalliser media |
| Simple payback (years) | 2–3 (avoided discharge fines only) | 3–5 (avoided fines + reduced makeup) | 4–6 (30–50% raw-water cut) | 6–9 (water-positive ESG only) |
| Permit costs (one-time) | $4K–$10K | $4K–$10K | $4K–$10K | $4K–$10K |
Simple payback improves to 3–5 years once avoided UzNature fines, avoided municipal tariff increases, and reduced makeup are counted against Option B. Option C captures a 30–50% raw-water cut without the full ZLD CAPEX burden, and it remains the strongest cost-to-ESG ratio for most Tashkent sites unless water-positive procurement language is contractually binding. For a Nordic reference on cold-climate blowdown economics, see the Stockholm data center cooling blowdown guide.
Frequently Asked Questions
What cycles of concentration can a Tashkent cooling tower run at with side-stream softening?
With side-stream softening or anti-scalant, a Tashkent cooling tower can run at 4–6 cycles of concentration before silica or CaCO₃ scaling forces blowdown. The blowdown-ratio math is non-linear: at 4 CoC, blowdown equals 25% of makeup; at 6 CoC, it drops to 20%, a 5 percentage-point reduction that delivers roughly 20% less blowdown volume, not 50% (per Genesis Water Tech 2025).
Is zero liquid discharge mandatory for data centers in Uzbekistan as of 2026?
No nationwide ZLD mandate exists as of 2026. Uzbekistan is not yet basin-classified as water-stressed at the federal level, but Tashkent sits in a transboundary Chirchiq watershed with declining per-capita renewable resource, and sovereign-cloud procurement is already writing water-positive language into contracts. Reserve RO preconcentrator + MVC + crystalliser for sites where municipal discharge is fully prohibited or where ESG posture demands closed-loop (per HydropureWater Baku 2026 ZLD bench, applied to Uzbekistan's permit frame).
How long does the Uzbekistan permit process take for a 5–10 MW data center?
Total permit timeline 6–10 months across the UzNature EIA and the Goswatercadastre / water-resources consent. UzNature EIA scoping and submission runs 60–120 days; the Goswatercadastre consent follows once the EIA is cleared. Running the two submissions in parallel rather than serially is the single biggest schedule lever, and the chemical dosing skid and sludge plan must be locked at the UzNature design-submission step, not at final inspection (per HydropureWater Astana 2026 MENR analog, applied to Uzbekistan's environmental code structure 2025).
Why can't sanitary sewage and cooling-tower blowdown share one clarifier?
Sanitary sewage is organic, low-flow, and pathogen-bearing at BOD₅ 200–300 mg/L; cooling-tower blowdown is mineralised, warm, and biocide-bearing at TDS 2,400–6,400 mg/L. Combining them overloads the biology, ruins the reuse chemistry, and pushes pathogen counts past reuse limits. Treat them on parallel trains — an A/O package plant for sanitary sewage and a DAF-to-RO train for blowdown — and only blend at the irrigation reuse point if both streams independently meet the SanPiD-equivalent target of fecal coliform ≤1,000 CFU/100 mL, BOD ≤50 mg/L, and turbidity <1 NTU.
