Data Center Wastewater & Cooling Blowdown Treatment in Tbilisi, Georgia (2026 Guide)

Why a Tbilisi data center breaks the U.S. data-center playbook

A data center in Tbilisi, Georgia needs a four-part treatment envelope: (1) feed-water conditioning sized to Tbilisi's 22–24 °C peak wet-bulb summers, (2) cycles-of-concentration optimization that pushes blowdown from 25% to ~20% of makeup, (3) a blowdown polishing train — side-stream filtration plus softening, brackish RO, or ZLD — that hits either Mtkvari discharge limits or Tbilisi Water utility sewer local limits, and (4) a closed-loop or hybrid cooling retrofit to capture the 50–70% freshwater savings that open evaporative towers forfeit in the Kura basin.

Tbilisi sits in a continental/semi-arid envelope where peak summer wet-bulb lands in the 22–24 °C band. That envelope matters because evaporative loss from a cooling tower scales with the wet-bulb depression between the air and the recirculating water, so a Tbilisi summer tower loses more makeup per ton-hour of cooling than the Atlanta baseline that most procurement decks reference. To hold the same makeup-to-blowdown ratio under harder working conditions, the cycle count has to be pushed harder, which in turn raises the TDS, hardness and silica load that the discharge train must handle. The Mtkvari (Kura) basin supplies a large share of Tbilisi's drinking water after treatment at the Gardabani and surface-water complexes, and 80–90% of U.S. data-center withdrawal sits in watersheds shared with public water systems (UGA TP-121, 2026-05). Tbilisi's exposure is at least as tight, because the same river supplies both the data center and the population it sits inside.

The regulatory stack is not the U.S. NPDES model. Ambient surface-water norms are set by the Ministry of Environmental Protection and Agriculture under Georgia's 2023–2026 water-law reform, and discharge to the Tbilisi Water utility sewer is governed by the utility's industrial-discharge rules. The Georgia U.S. state analogue still helps anchor the conversation: 213 operational data centers as of 20 May 2026, behind Virginia (603), Texas (461), California (287) and Illinois (228) (UGA TP-121, 2026-05). No site-specific Tbilisi capacity figure sits in the public research, so the design envelope should be planned across a 1–10 MW range, with 10 MW used as the worked example in the closing section. For a U.S. baseline comparison that the procurement team can carry into a Tbilisi meeting, see HydropureWater's Atlanta data-center cooling-blowdown guide.

What cooling-tower blowdown actually contains in this climate

Cooling-tower feed in the regional baseline carries 40–80 mg/L total hardness and 30–60 mg/L silica (HydropureWater field data, 2026). Tbilisi's harder groundwater and Mtkvari-derived supply sit at or above the upper end of that band, and the continental/semi-arid evaporation load pulls blowdown concentrations even higher. At 4 cycles of concentration, blowdown reaches 250–500 mg/L hardness, 150–300 mg/L silica, 500–1,500 mg/L chloride, and TDS values above 2,000 mg/L before chemical additives are counted (HydropureWater field data, 2026). The scaling tendency is dominated by silica and calcium carbonate, the corrosion tendency by chloride and the corrosion metals (Zn, Cu, Pb) stripped from coil surfaces over time.

Blowdown ratio is governed by the 1/(CoC−1) relationship: 4 CoC gives 25% blowdown, 5 CoC gives 20%, 6 CoC gives ~20% of makeup as a controlled purge (Genesis Water Tech, 2026). Tbilisi's higher makeup demand from evaporative loss magnifies the absolute volume in all three cases, so a 1 percentage-point blowdown reduction saves more cubic metres per day in Tbilisi than it does in Atlanta. Cooling-tower chemistry also carries biocides, scale inhibitors and — for any site running direct-liquid or immersion cooling — PFAS monitoring triggers that the receiving utility or surface water will see at the discharge point. The parameter envelope the Tbilisi utility or Mtkvari target is most likely to focus on mirrors the U.S. local-limit pattern: TDS, TSS, total metals, pH 5.0–10.0, oil & grease, and temperature. Hyperscalers in water-stressed regions are already moving toward on-site reuse rather than sewer or surface-water paths; the Guayaquil data-center cooling-blowdown guide covers a comparable tropical-basin case for reference.

Three discharge paths, one decision

The first procurement-side decision is not which treatment train to buy; it is where the blowdown will go. The choice locks in the treatment envelope, the permit timeline and the disclosure story.

Path 1 — Direct discharge to Mtkvari or a tributary. This path requires an ambient surface-water permit under Georgia's water-law framework, with temperature and dissolved-oxygen scrutiny analogous to U.S. Clean Water Act §316 thermal-plume review. The U.S. analogue is NPDES permits for surface-water discharges, which Water Tech Online reports as state-issued and typically taking 18 months or more to obtain (2026-09-11). A Tbilisi permit timeline is not in the public research, but the structural risk of a long-lead environmental review should be planned for at RFP stage.

Path 2 — Sanitary sewer to Tbilisi Water. Industrial discharge approval is needed, and the utility's local limits will be the binding envelope, with TDS, pH, metals and temperature as the most likely pressure points. This path avoids the surface-water permit but trades it for utility local-limit risk and sewer fees that scale with load.

Path 3 — On-site reuse. Cooling-tower makeup, adiabatic-assist pre-cooling, landscape or irrigation. This collapses the discharge question into a treatment-train question, and is the path the EPA Water Reuse Action Plan 2.0 endorses for water-stressed sites. Google's Douglas County operation — where treated reclaimed wastewater from the county water and sewer authority is used for data-center cooling — is the closest public analogue (EPA Water Reuse Action Plan 2.0, via Water Tech Online, 2026-09-11). It is also the direction hyperscalers are moving because it is the most disclosure-defensible on a water-stewardship report.

Four process trains, ranked by discharge liability

The four canonical process options cover every discharge scenario a Tbilisi site will face, and they scale with discharge liability and water-stewardship value.

Side-stream multi-media filter plus weak-acid cation or lime-soda softening. This is the entry-level train for any Tbilisi site. It lifts cycles from 4 to 6, cuts blowdown ~30% (HydropureWater field data, 2026), and lands in the $50,000–$200,000 CAPEX band (Genesis Water Tech, 2026). It unlocks both the Mtkvari and the Tbilisi Water sewer paths because the softened blowdown stays below the local-limit envelope most utilities apply. See the side-stream softening reference and the side-stream multi-media filtration train for equipment context.

Brackish reverse osmosis on cooling-tower blowdown. RO recovery on cooling-tower blowdown realistically operates at 50–85%, with permeate at 10–50 mg/L TDS returning as cooling-tower makeup and concentrate going to sewer or to a downstream concentrator (Genesis Water Tech, 2026). This is the lowest-cost path that pairs with a water-stewardship disclosure, which is the strategic direction hyperscalers are moving. The recovery ceiling is set by silica scaling on the concentrate side, so softening upstream is normally a prerequisite. For a head-to-head against the alternatives, see the brackish RO vs alternatives comparison and the brackish RO on cooling-tower blowdown product reference.

Full zero-liquid discharge with brine concentrator and crystallizer. 95–99% overall recovery, solid waste under 1% of original blowdown volume, $3–8M CAPEX for typical data center capacity (Genesis Water Tech, 2026). ZLD is justified where both the Mtkvari and sewer paths are closed and water-stewardship disclosure is a hard requirement. Energy intensity is the trade-off.

Closed-loop or hybrid cooling retrofit. World Economic Forum data cited by Water Tech Online (2026-09-11) puts closed-loop water reduction above 50% versus open evaporative, and UGA TP-121 frames the upper end at up to 70% (2026-05). The trade-off is more complex piping and higher electricity for compressors and pumps. In a continental/semi-arid envelope where the freshwater forfeit of open evaporative is at its worst, the case for closed-loop or adiabatic-assist hybrid is strongest.

Worked example: 10 MW Tbilisi site, open evaporative tower, 6 CoC + RO

Anchor the math to two published reference points. A 10 MW facility using evaporative cooling at 4 CoC can intake on the order of 15 million gallons monthly, of which 25% — 3.75 million gallons — leaves as recoverable blowdown (Genesis Water Tech, 2026). Pushing cycles from 4 to 6 drops the blowdown share to ~20% of makeup, a 5-percentage-point absolute reduction. The side-stream filtration and softening package at the entry-level CAPEX of $50,000–$200,000 (Genesis Water Tech, 2026) makes that 6 CoC operation realistic without breaching silica scale limits.

For the RO train, the reference example is 60,000 gpd blowdown at 65% RO recovery: permeate ≈ 39,000 gpd, concentrate ≈ 21,000 gpd (HydropureWater). In SI units, 750,000 gpd — the order-of-magnitude monthly blowdown for the worked 10 MW site at 6 CoC — converts to roughly 2,840 m³/day at the 1 m³ = 264.17 gal factor. The RO permeate returned to cooling-tower makeup on that envelope lands near 1,850 m³/day, with concentrate near 990 m³/day routed to sewer or to a downstream concentrator depending on the discharge path chosen. The RO train CAPEX is site-specific and not published in the public research, so the sizing logic is what the engineer should carry into procurement rather than a fabricated dollar figure. The combined envelope — softening at $50,000–$200,000 plus the RO train sized to the 60,000 gpd order of magnitude — gives a procurement-defensible CAPEX band that the committee can stress-test against flow.

Payback framing: cycles 4→6 plus 65% RO on a 10 MW-class facility gives roughly 25% freshwater reduction (HydropureWater field data, 2026), supporting a 3–5 year payback under typical water-and-sewer rates. The local Tbilisi rate context is qualitative in this article because no public rate figure is in the research, but the freshwater-and-sewer savings line on a hyperscaler P&L is the same shape it is in Atlanta.

Climate, water-source, and disclosure pressures specific to Tbilisi

Closed-loop and adiabatic-assist cooling become more attractive in a continental/semi-arid envelope because the freshwater forfeit of open evaporative is highest when peak wet-bulb is highest. Tbilisi's 22–24 °C summer wet-bulb sits at the upper end of the operating window where the energy penalty of closed-loop is easiest to justify. The Tbilisi Water utility and Mtkvari ambient norms are the binding discharge constraints; the U.S. analogue — 40 CFR Part 403 categorical and local limits, state-issued NPDES — is a useful benchmark for the structure but not the controlling framework, and a Tbilisi-specific permit timeline is not in the public research.

Disclosure pressure is the second-order driver. Water Tech Online's September 2026 feature catalogues more than 100 local U.S. construction moratoriums and a public-opinion landscape in which roughly half of Americans opposed to nearby data centers flag water as their top concern (2026-09-11). Tbilisi-specific polling is not in the research, but the political shape of the conversation is the same — a project designed now should be ready to claim reuse and freshwater-reduction credit publicly. The May 2026 introduction of the U.S. Advancing Water Reuse Act by Senators Luján and Britt, which would create federal tax credits for on-site recycling and municipal water-reuse upgrades, is a leading indicator that on-site recycling will move from voluntary to incentivized (Water Tech Online, 2026-09-11). Any Tbilisi project designed in 2026 should be built so it can claim a comparable credit under whatever reuse instrument the Georgian policy track eventually produces.

Frequently Asked Questions

What wastewater and cooling blowdown treatment does a data center in Tbilisi, Georgia need?

A Tbilisi data center needs a four-part envelope: feed-water conditioning sized to 22–24 °C peak wet-bulb summers, cycles-of-concentration optimization that pushes blowdown from 25% to ~20% of makeup, a blowdown polishing train (side-stream filtration plus softening, brackish RO, or ZLD) that hits Mtkvari or Tbilisi Water sewer limits, and a closed-loop or hybrid cooling retrofit that captures the 50–70% freshwater savings open evaporative towers forfeit in the Kura basin.

Which discharge path should a Tbilisi data center use for cooling-tower blowdown?

Three paths exist — direct discharge to Mtkvari under Ministry of Environmental Protection and Agriculture ambient norms, sanitary sewer to Tbilisi Water under utility local limits, and on-site reuse. The on-site reuse path is the most disclosure-defensible because it collapses the permit question into a treatment-train question, and it is the direction the EPA Water Reuse Action Plan 2.0 endorses for water-stressed sites. Mtkvari and sewer remain viable but each carries its own lead-time or local-limit risk.

How much does cooling-tower blowdown treatment cost in 2026?

Side-stream filtration plus softening lands in the $50,000–$200,000 CAPEX band and supports 6 cycles of concentration with ~30% blowdown reduction (Genesis Water Tech, 2026). Brackish RO on the blowdown stream operates at 50–85% recovery and is sized to site-specific flow. Full ZLD with brine concentrator and crystallizer achieves 95–99% recovery but costs $3–8 million for typical data-center capacity, and is justified only where both surface-water and sewer paths are closed.

Is closed-loop cooling worth the energy penalty in Tbilisi's climate?

Closed-loop and hybrid designs cut water consumption by 50–70% versus open evaporative cooling (UGA TP-121, 2026-05; Water Tech Online, 2026-09-11), at the cost of higher electricity use for compressors and pumps and more complex piping. In a continental/semi-arid envelope where peak wet-bulb reaches 22–24 °C and the freshwater forfeit of open evaporative is at its worst, the water-energy trade-off usually favours closed-loop or adiabatic-assist hybrid.

What is the typical Tbilisi data center water consumption per MW?

No Tbilisi-specific per-MW figure sits in the public research, so the design envelope should be planned across a 1–10 MW range. As an order-of-magnitude reference, a 10 MW facility using evaporative cooling at 4 cycles of concentration can intake on the order of 15 million gallons monthly, with 25% — 3.75 million gallons — leaving as recoverable blowdown (Genesis Water Tech, 2026). Pushing cycles from 4 to 6 drops the blowdown share to ~20% of makeup, a 5-percentage-point absolute gain.