Why the Pavlodar basin and the Ekibastuz Data Center Valley make reuse a 2026 permit issue
Kazakhstan's Ministry of Artificial Intelligence and Digital Development has confirmed a Data Center Valley pilot site in Ekibastuz on a 200-hectare land plot in the Pavlodar Region, with phased expansion from an initial 300 MW toward 1 GW, roughly $30 billion in projected investment, and KT-Telecom as the operator of record (Astana Times, 7 April 2026). Network latency from Ekibastuz is estimated at 80–88 ms to Central Asian hubs, materially below the 120–150 ms typical of Central Asia–Frankfurt/Hong Kong routes, which is the design rationale for siting hyperscale capacity on the central-Irtysh steppe (Astana Times, 7 April 2026). The water consequence is that the Irtysh/Pavlodar system already allocates to Ekibastuz GRES-1 and GRES-2 power generation, Pavlodar municipal demand, and existing industrial users, so a hyperscale data hall or co-located fab entering the basin in 2026 is reviewed as a closed-basin applicant rather than a discretionary one.
The global context is unforgiving: 40% of existing fabs and over 40% of new fabs announced since 2021 are projected to sit in basins with high or extremely high water-stress risk by 2030, and 45% of data centres globally already sit in river basins at high water-availability risk (Lepawsky 2024 and Hajonides et al. 2025, via TNFD, Feb 2026). The comparable closed-basin jurisdictions in 2026 — Baku/Caspian, Yerevan/Ararat, Isfahan/Zayandeh Rud — anchor an 80% internal-reuse floor as a permit pre-condition rather than a corporate ESG preference (HydropureWater Baku 2026 guide; HydropureWater Isfahan 2026 guide). For a Pavlodar project, that floor should be briefed to MENR and the Kazakhstan Ecology Committee in the kickoff meeting, not introduced at the EIA submission stage, because the consent clock otherwise resets. The parallel engineering template in the Isfahan 2026 guide to fab and data-hall process wastewater walks the same allocation-closure arithmetic for the Zayandeh Rud basin and is the closest design analogue outside the Caspian/Ararat pair.
Separating the two wastewater streams: fab wet drains versus data-hall cooling-tower blowdown
Forcing fab wet-process drains and data-hall cooling-tower blowdown (CTBD) through a single neutralization/precipitation stage is the most common reason a 2026 ETP gets over-sized in northern Kazakhstan. The two streams are chemically different, and blending them forces the chemistry to be sized for the upset case rather than the average case.
Data-hall CTBD is dominated by TDS, Ca/Mg hardness, silica, and trace oxidizing biocides, typically running at 4–6 cycles of concentration (COC) with TDS up to 2,000 ppm at 30–40 °C (HydropureWater Baku 2026 guide). AHU condensate typically sits below 50 mg/L TDS and should be segregated on its own line because glycol from coil leaks requires stripping, not blending into the main RO loop. Fab wet-process drains carry HF-etch fluoride at 50–500 mg/L, TMAH from the developer stream, CMP nanoparticles (silica, ceria, alumina), ammonia, and mixed acids/alkalis; Cu can reach 100 mg/L in untreated CMP effluent (Lai & Lin 2004, via HydropureWater Isfahan 2026 guide). The two envelopes do not belong in the same equalization tank.
The defensible 2026 architecture is segregated drains at source, separate equalization, and only a final common RO or brine-concentrator polishing step if upstream chemistries are genuinely compatible. Municipal-sewer discharge to Astana or Pavlodar WWTPs is not a 2026 design option at hyperscale flow because the receiving works are not sized for fab chemistry or for CTBD volumes in the 200–1,000 m³/day band, which mirrors the Baku Caspian/Maraza constraint (HydropureWater Baku 2026 guide).
| Parameter | Fab wet-process drain | Data-hall CTBD |
|---|---|---|
| Dominant species | HF-etch fluoride (50–500 mg/L), TMAH, CMP nanoparticles, Cu up to 100 mg/L, ammonia, mixed acids/alkalis | TDS, Ca/Mg hardness, silica, trace oxidizing biocides |
| pH behaviour | Strongly acidic to strongly alkaline, highly variable | Near-neutral, buffered by treatment chemistry |
| Suspended load | Abrasive CMP nanoparticles, precipitated metal hydroxides | Mostly dissolved; minor suspended scale |
| Design objective | Toxicity removal + high-recovery reuse | Scale and silica control + cooling-tower make-up |
| 2026 unit-process answer | Precipitation + UF + RO + MVC brine concentration (full-stream ZLD when HF, TMAH, nanoparticles co-occur) | Softening + side-stream RO; brine haul-off or small MVC on RO brine only |
The 2026 treatment train: MBR, two-pass RO, and where MVC earns its place
The unit-process sequence below is what holds up in front of a MENR or Kazakhstan Ecology Committee reviewer in 2026 and aligns with the train documented in the Baku and Isfahan 2026 guides for comparable closed-basin jurisdictions.
Stage 1 — Segregation and equalization. A dedicated EQ tank with 4–8 h HRT and online pH/conductivity dampens the 1–5 pH excursions that follow chiller trips; AHU condensate is segregated on its own line.
Stage 2 — DAF and multi-media filtration. A DAF unit in the 4–300 m³/h class floats oils, biofilm, and metal-hydroxide floc, followed by a multi-media filter that drops SDI below 3 and protects the RO from Cu, Fe, and Zn fouling. For fab trains with Cu-CMP load, a hollow-fiber UF pretreatment upstream of DAF improves colloid removal and stabilizes downstream RO performance.
Stage 3 — Softening and antiscalant dosing. A twin-tank industrial softener in the 1–45 T/h class targets hardness below 50 mg/L as CaCO₃ and silica below 10 mg/L as SiO₂, with PLC-controlled antiscalant dosing tied to RO feed flow. On the variable Irtysh intake, the softener setpoint should track seasonal silica swings rather than hold a fixed value.
Stage 4 — MBR polishing (stand-alone halls only). A submerged PVDF MBR system with 0.1 µm membranes delivers below 1 NTU and below 10 mg/L COD, allowing direct RO feed where sanitary load is co-mingled. For fab streams, MBR is replaced by a dedicated HF-removal and metals-precipitation step upstream of UF. The module selection criteria that govern this step are covered in the 2026 MBR membrane module design criteria reference.
Stage 5 — Two-pass RO at 80–95% recovery. An industrial RO system with recovery rates up to 95% runs the first pass at 150–250 psi (10–17 bar) for bulk salts, with a second pass polishing to TDS below 200 mg/L and Cl⁻ below 100 mg/L. Above 95% recovery, silica scaling on the second-pass membranes drives CIP frequency up sharply; 80–95% is the practical operating window.
Stage 6 — MVC brine concentration. Mechanical vapor recompression at 25–40 kWh/m³ of brine concentrated is applied to the RO brine stream only for stand-alone halls, and to the full stream when HF, TMAH, and CMP nanoparticles co-occur (HydropureWater Baku 2026 guide).
Co-located fab versus stand-alone hall: the decision matrix that drives ZLD
The choice between high-recovery RO with brine haul-off and full-stream ZLD is a chemistry decision read off the influent, not a benchmark preference. Whenever HF-etch fluoride (50–500 mg/L), CMP nanoparticles, and TMAH co-occur in the wastewater envelope, full-stream ZLD is the 2026 default because the combined load cannot be diluted into a discharge consent (HydropureWater Baku 2026 guide). Stand-alone data-hall CTBD does not need full-stream ZLD; co-located fab-hall hybrids must default to it.
Samsung's published practice — automated process wastewater treatment with a central control room, continuous monitoring of the full purification process, three-stage containment covering wastewater inflow, treatment, and discharge, and backup treatment facilities across all production lines — is the operational template a MENR reviewer will expect for any co-located 2026 fab (Samsung Semiconductor sustainability page, accessed 2026).
| Project type | Influent envelope | 2026 unit-process answer | ZLD scope |
|---|---|---|---|
| Stand-alone hyperscale data hall (Astana or Ekibastuz, no fab) | CTBD only: TDS, hardness, silica, biocides | MBR + two-pass RO at 80–95% recovery | MVC on RO brine only, when Irtysh/Pavlodar discharge path is restricted |
| Co-located fab-hall hybrid | Fab wet drains + hall CTBD: HF (50–500 mg/L), TMAH, CMP nanoparticles, Cu up to 100 mg/L | Precipitation + UF + two-pass RO + MVC | Full-stream ZLD, non-negotiable baseline |
| Small OSAT or 150/200 mm legacy line | Fab wet drains at lower volumetric flow | Two-pass RO at 95% recovery + brine bleed | Brine to authorized evaporation or hazardous-waste disposal; economic optimum when ZLD energy cost is hard to justify |
Cold-climate deltas for Ekibastuz: winter design choices the Caspian and Yerevan guides skip
The Pavlodar/Irtysh corridor runs sub-zero in winter, and the operating envelope that condition imposes is not covered by the Baku/Caspian or Yerevan/Ararat templates. The four items below must be priced into the 2026 EPC scope rather than added as a site-installation afterthought.
Sub-zero winter ambient requires enclosed or heated RO rooms, glycol trace heating on RO feed lines, and housing for the DAF and MBR stages to keep biological activity above 10 °C. Cooling-tower operation in winter allows higher cycles of concentration (6–7 COC) because evaporative loss drops; the blowdown formula B = E/(COC−1) means the RO train size and the downstream brine volume both shrink as COC rises (HydropureWater Isfahan 2026 guide). Air-cooled adiabatic or dry-cooled hybrid loops are the cheapest ZLD-style answer for a hyperscale Ekibastuz hall on the central-Irtysh steppe, because evaporative loss collapses and the wastewater envelope reduces to AHU condensate plus a small blowdown side-stream. Freeze protection on brine lines and outdoor chemical dosing tanks must be in the 2026 CAPEX band — confirm with the vendor that dosing skids and chlorine dioxide generators are factory-winterized for sub-zero shipment.
2026 CAPEX and OPEX bands, and the avoided-discharge payback
The bands below are 2026 engineering estimates, not firm quotes; a vendor proposal will move the band based on metallurgy, automation scope, and ZLD inclusion (HydropureWater Baku 2026 guide; HydropureWater Isfahan 2026 guide).
| Project size | 2026 CAPEX band (USD/m³/day installed) | Dominant OPEX lines |
|---|---|---|
| Small hall, package plant + haul-off | $150–300 | Brine haul-off, softener regeneration |
| Mid-size 200–1,000 m³/day MBR + RO (typical 5–20 MW Ekibastuz hall) | $400–700 | RO CIP, antiscalant, ClO₂ residual |
| Hyperscale or co-located fab with brine ZLD | $800–1,200 | MVC electricity 25–40 kWh/m³, high-silica RO replacement |
OPEX lines that push cost up: membrane CIP frequency above 95% recovery on the second pass due to silica scaling; isothiazolinone biocide residuals shortening RO life, which is the case for a ClO₂ side-loop upstream of the RO rather than dosing in the tower; MVC electricity at 25–40 kWh/m³ of brine concentrated (HydropureWater Baku 2026 guide). Avoided-discharge math at $5–15/kgal: 100 m³/day of untreated blowdown at the upper end is roughly USD 400/day, so an 80% recovery RO pays back inside about 24 months at hyperscale flow. Request a line-item split of ZLD thermal equipment versus membrane equipment in any vendor proposal, and confirm that high-silica-rated RO and UF elements are specified, not generic elements. Where a multimedia filter sits upstream of the RO, confirm the multi-media filter specification matches the seasonal Irtysh silica swing rather than a single feed point.
Frequently Asked Questions
What reuse percentage does MENR or the Kazakhstan Ecology Committee expect from a 2026 Astana or Ekibastuz fab or data hall?
The practical 2026 floor is 80% internal reuse, mirroring the closed-basin practice in Baku/Caspian, Yerevan/Ararat, and Isfahan/Zayandeh Rud, where 80% is anchored as a permit pre-condition rather than a corporate ESG preference (HydropureWater Baku 2026 guide; HydropureWater Isfahan 2026 guide). The mass balance should be designed around 80–95% recovery from day one so the consent clock does not reset at EIA submission.
What 2026 CAPEX should a buyer budget for a 5–20 MW Ekibastuz hyperscale data hall?
A mid-size 200–1,000 m³/day MBR + RO train typically sits at $400–700/m³/day installed, which is where the 5–20 MW Ekibastuz hall scope lands in 2026; payback on the 80% recovery RO is inside approximately 24 months at hyperscale flow on the avoided-discharge math of $5–15/kgal (HydropureWater Baku 2026 guide). Final pricing depends on metallurgy, automation scope, and whether MVC is in or out of the package, so request a line-item split of ZLD thermal equipment versus membrane equipment in any vendor proposal.
Can a 2026 Ekibastuz project discharge to the municipal sewer at hyperscale flow?
No. The Astana and Pavlodar WWTPs are not sized for fab chemistry or for CTBD volumes in the 200–1,000 m³/day band, and MENR treats municipal-sewer discharge as a non-starter for ICT projects above the small-hall threshold in 2026, mirroring the Baku Caspian/Maraza constraint (HydropureWater Baku 2026 guide).
What vendor documentation should a buyer require for a 2026 Pavlodar fab or hyperscale hall?
Require OCEMS-ready pH/conductivity/TOC/fluoride monitoring, ISO 14001 certification, a real Kazakhstan or Central Asia reference list, an RO energy-recovery device rated for sustained high-recovery operation, and winterized automatic chemical dosing skids for sub-zero ambient. Confirm that the UF membrane is rated for the fluoride and solvent excursions the equalization tank can deliver, and that RO elements are rated for the high-silica Irtysh feed rather than a generic specification.
