The 2026 Budapest site map: three streams, not one
Treat fab process wastewater as three hydraulically separate streams at a Budapest site: cooling-tower blowdown (CTBD), fab process/condensate, and sanitary/facility waste. Stream 1 (CTBD) typically runs 1,200–6,000 mg/L TDS, 10–50 mg/L TSS, dominated by calcium hardness and silica, with residual ClO₂ and isothiazolinone biocides, phosphonate scale inhibitors, and trace Fe/Cu from corrosion. Stream 2 (fab process/condensate) carries UPW reject with fluoride often >50 mg/L as F from HF/SC1/SC2 etches, ammonia at 50–500 mg/L as N, dissolved silica, and CMP slurry carryover at low pH with 20–200 mg/L TSS and Cu/Ni/Co at 0.1–10 mg/L each. Stream 3 (sanitary/facility) is conventional domestic-strength, routed through a packaged MBR held independent of cooling chemistry, and is the only stream that collapses when the site is data-hall only.
Stream separation is not optional. Blending untreated blowdown into fab reject loads the RO with biofouling precursors and destabilises fluoride rejection, because biofouling compresses the membrane's effective divalent-rejection surface area. The macro picture makes the engineering case before any pipe is sized: globally, semiconductor water use doubled between 2012 and 2022; a single fab uses around 14 billion litres of UPW per year; and 1.4–1.6 units of municipal water are needed per 1 unit of UPW (TNFD, Feb 2026, citing WEF 2025 and IDE Technologies 2024). For Budapest planning, where both fabs and 5–50 MW data halls are scaling into the same Danube basin, keeping these three streams segregated is the only way to keep the IED 2.0 envelope, the cooling reuse target, and the FCM/OKF permit all simultaneously satisfiable.
| Stream | Key parameters | Dominant chemistry | Default conditioning |
|---|---|---|---|
| CTBD | 1,200–6,000 mg/L TDS, 10–50 mg/L TSS | Ca/silica hardness, ClO₂/isothiazolinone, phosphonates, trace Fe/Cu | DAF or multimedia filter + UF; biocide quench |
| Fab process/condensate | F >50 mg/L, NH₃-N 50–500 mg/L, TSS 20–200 mg/L, Cu/Ni/Co 0.1–10 mg/L | HF/SC1/SC2 etches, CMP slurry, dissolved silica | CaCl₂/Ca(OH)₂ precipitation + lamella clarifier + UF |
| Sanitary/facility | Domestic-strength | BOD/COD/ammonia, no cooling or etch chemistry | Packaged MBR, independent of cooling loop |
The 2026 compliance stack that binds a Budapest site
IED 2.0 BAT-AELs for semiconductor surface treatment set numeric limits on fluoride, total nitrogen, and heavy metals that bind regardless of whether the final discharge point is the municipal sewer (FCSM) or a Danube-tributary surface-water body; this is the framing any 2026 Budapest submission has to satisfy on both pathways at once. UWWTD 91/271/EEC, with the 2024 amendment, governs indirect-discharge pretreatment thresholds, and the 2024 amendment is the operative 2026 text reviewers will apply to new Budapest submissions.
In Hungary, the transposition layer is Government Decree 309/2014 (Korm. rendelet) on BAT and emission limit values, with Decree 220/2004 setting BAT conclusions for waste-water-intensive industries. Site-specific industrial discharge permits run through the FCM/OKF (Fogyasztóvédelmi / Katasztrófavédelmi) pipeline at municipal/county level, functionally the Hungarian analogue to the HOFOR/Miljøstyrelsen permit stack used in the Copenhagen reference frame. EED 2023/1791 requires annual waste-heat and water cost-benefit reporting for data centres above 1 MW from October 2025 and sets a PUE ≤1.2 target live in 2026, which moves reuse from optional to baseline. The IED 2.0 surface-treatment BAT-AEL stack does not apply to a data-hall-only site, but LCP BREF triggers if on-site combustion exceeds 50 MWth, which is the typical N+1 diesel envelope at a 20 MW IT hall.
| Instrument | Scope in Budapest | 2026 binding effect |
|---|---|---|
| IED 2.0 recast BAT-AELs (surface treatment) | Fab co-locate, any discharge path | Numeric F, total N, heavy metals bind |
| UWWTD 91/271/EEC + 2024 amend. | Indirect discharge to sewer | Pretreatment thresholds apply to FCSM |
| Government Decree 309/2014 | Hungarian BAT transposition | Sets national ELVs and BAT obligations |
| Decree 220/2004 | WW-intensive industries | BAT conclusions referenced in permit |
| FCM/OKF industrial discharge permit | Municipal/county issue | Site-specific ELVs and monitoring |
| EED 2023/1791 | Data centres >1 MW | Reporting from Oct 2025; PUE ≤1.2 target |
| LCP BREF | Combustion >50 MWth | Triggers at N+1 diesel block of 20 MW hall |
Stage 1 chemistry and Stage 2 RO: the 2026 default train
Stage 1 conditions each stream separately upstream of any combined unit operation. For fab UPW reject, pH adjustment with CaCl₂ or Ca(OH)₂ precipitates fluoride to <15 mg/L and drops total metals to <1 mg/L; a high-efficiency lamella clarifier handles the high-solids CMP slurry load, and a 0.03 µm PVDF ultrafiltration skid polishes the clarifier overflow to TSS <10 mg/L and turbidity <1 NTU at the RO feed (HydropureWater field data, 2026). On the CTBD side, a DAF unit or side-stream multimedia filter strips TSS and oil carryover; downstream UF polishes to TSS <5 mg/L and oil <2 mg/L; biocide residual is quenched ahead of RO to protect thin-film composite membranes (HydropureWater field data, 2026). A PLC-controlled chemical dosing skid holds reagent stoichiometry tight on both streams and prevents overdosing from inflating both OPEX and downstream fouling.
Stage 2 is a brackish-water industrial RO membrane system with up to 95% recovery on the combined feed. The operating setpoint sits at 70–85% depending on feed TDS, with the lower-TDS blowdown fraction used to lift overall recovery by diluting the fab reject's fluoride and silica load. Permeate TDS is held <50 mg/L, suitable for cooling-tower makeup at 50–70% reuse; an optional MBR polish upstream drops TOC and ammonia where biological loads persist, ahead of the thin-film composite membranes. The full Stage 1 + Stage 2 train has been demonstrated at 99.8% contaminant removal on comparable high-purity reuse feeds (HydropureWater field data, 2026), with fluoride <5 mg/L, total metals <0.5 mg/L, and concentrate TDS <1,500 mg/L at the discharge point, comfortably inside both the FCSM sewer envelope and the IED 2.0 BAT-AEL binding numeric limits.
| Stage | Unit | Setpoint / target | Purpose |
|---|---|---|---|
| 1a — Fab reject | CaCl₂ / Ca(OH)₂ precipitation + high-efficiency lamella clarifier | F <15 mg/L, metals <1 mg/L total | BAT-AEL feed condition; CMP slurry drop-out |
| 1b — CTBD | DAF unit or multimedia + UF | TSS <5 mg/L, oil <2 mg/L | RO membrane protection, biocide quench |
| 1c — UF polish | 0.03 µm PVDF ultrafiltration skid | TSS <10 mg/L, turbidity <1 NTU | Final RO feed guard |
| 1d — Reagent control | PLC-controlled chemical dosing skid | Stoichiometry on both streams | Limits OPEX and downstream fouling |
| 2 — RO | Industrial RO membrane system | 70–85% recovery, permeate <50 mg/L TDS | Cooling-tower makeup at 50–70% reuse |
Posture A, B or C: which Budapest discharge path to size for
Posture A is direct sewer discharge under an FCM/OKF industrial permit: lowest CAPEX, highest water-stress exposure because the permit can be tightened during a Danube-basin summer low-flow event; it fits edge halls under ~5 MW IT with short payback pressure. Posture B is high-recovery RO plus concentrate disposal to sewer or solid waste, the 2026 default for most Budapest co-located sites, demonstrated at 99.8% removal with 50–70% cooling reuse (HydropureWater field data, 2026). Posture C is hybrid ZLD: RO plus thermal evaporator or crystalliser, only worth the CAPEX premium if the site sits in a declared water-emergency zone or a corporate net-zero water target is binding.
Three decision drivers should be weighed explicitly: the FCM/OKF sewer tariff band; the evaporator steam tariff if ZLD is chosen; and the value of the EED 2023/1791 waste-heat credit where the concentrate evaporator is heat-integrated with a district-heating export loop. For sizing, a Nordic 22.5 MW data hall sits at 400,000–500,000 L/day at an industry WUE band of 1.8–2.5 L/kWh, so a Budapest 22.5 MW air-cooled adiabatic or chilled-water hall is best engineered to that band before locking the train. Engineers comparing city overlays can cross-check against the Munich semiconductor engineering guide and the Frankfurt 2026 compliance guide for a colder-climate and Rhine-Main comparison frame.
| Posture | Configuration | Best fit in Budapest | Water-stress exposure |
|---|---|---|---|
| A — Direct sewer (FCM/OKF) | Conditioning only, sewer discharge | Edge halls <5 MW IT, short payback pressure | High — permit tightens in summer low-flow |
| B — High-recovery RO + concentrate disposal | Full Stage 1 + RO; brine to sewer or solid waste | Most Budapest co-located sites | Moderate — controlled by 99.8% envelope |
| C — Hybrid ZLD | RO + evaporator / crystalliser | Water-emergency zones, net-zero binding sites | Lowest — only if EED heat credit applies |
Cost envelope and PFAS overlay: the 2026 Budapest numbers
Indicative 2026 CAPEX for a Budapest build sits at EUR 0.8M–1.5M for a small/edge data hall (below 5 MW IT) and EUR 1.5M–3.5M for a hyperscale cluster above 20 MW IT, with the scaling drivers being permeate quality, reuse rate, and concentrate handling (HydropureWater field data, 2026). OPEX splits roughly 45% energy (pumps, RO high-pressure pump, optional evaporator), 25% membrane and media replacement, 20% chemical dosing, and 10% labour and monitoring; a 50–70% cooling-tower reuse target pays back the CAPEX in roughly 3–5 years on the water and sewerage line items alone, before EED 2023/1791 waste-heat credits are counted (HydropureWater field data, 2026).
The PFAS overlay is a 2026 procurement item, not an option: semiconductor PFAS discharges to wastewater are under active consortium survey in 2026, so any Budapest site using fluorinated chemistries in etch, CVD chamber clean, or fire-suppression should plan for a PFAS monitoring line item in the FCM/OKF permit renewal. Concentrate handling is the real 2026 design question, not the RO itself: thermal ZLD with crystalliser, high-recovery RO plus evaporator of residual brine, or crystalliser-only tied to a salt-cake disposal contract. For fluoride-specific chemistry economics, the engineering spec for HF wastewater treatment by chemical precipitation is a useful 2026 benchmark. For broader cost benchmarking on the CMP side, see the CMP wastewater equipment cost comparison.
| Item | Small/edge hall <5 MW IT | Hyperscale cluster >20 MW IT | Notes |
|---|---|---|---|
| CAPEX envelope | EUR 0.8M–1.5M | EUR 1.5M–3.5M | Driven by permeate quality, reuse %, concentrate handling |
| OPEX split | ~45% energy / 25% membranes / 20% chemicals / 10% labour | Same proportions at scale | Evaporator tariff dominates if ZLD chosen |
| Payback (water + sewerage only) | 3–5 years | 3–5 years | EED 2023/1791 heat credit shortens this |
| PFAS line item | Monitoring in FCM/OKF renewal | Monitoring + possible treatment | Consortium survey live in 2026 |
Frequently Asked Questions
How long does an FCM/OKF industrial discharge permit take for a Budapest fab-style submission in 2026?
Typical lead time is 6–12 months depending on the county/municipality route through which the FCM/OKF pipeline runs, by analogy with the HOFOR sub-system lead-time spread transposed to the Hungarian administrative frame (HydropureWater field data, 2026, transposed to Hungary). Buyers should request a written critical-path schedule from the permitting consultant before locking Stage 1 equipment delivery dates.
What 2026 CAPEX should a Budapest engineer budget for a 5–20 MW data hall with 50–70% cooling reuse?
For a hyperscale cluster above 20 MW IT, the indicative 2026 CAPEX envelope is EUR 1.5M–3.5M, with the scaling drivers being permeate quality, reuse rate, and concentrate handling (HydropureWater field data, 2026). OPEX splits roughly 45% energy, 25% membranes and media, 20% chemicals, and 10% labour. Because permit authority, sewer tariff, and evaporator steam tariff all move the number, the buyer must request a site-specific quote with feed-water analysis, target reuse percentage, and concentrate destination stated explicitly.
How do I choose between a direct sewer permit and a hybrid ZLD posture for a Budapest site?
Posture A (direct sewer under FCM/OKF) fits edge halls under ~5 MW IT with short payback pressure but carries the highest water-stress exposure, because the permit can tighten during a Danube-basin summer low-flow event. Posture B (high-recovery RO plus concentrate disposal) is the 2026 default for most co-located sites. Posture C (hybrid ZLD) is only worth the CAPEX premium if the site sits in a declared water-emergency zone or a corporate net-zero water target is binding. Decision inputs a buyer must obtain: the FCM/OKF sewer tariff, the evaporator steam tariff if ZLD is chosen, and the value of the EED 2023/1791 waste-heat credit at the specific site.
What is the demonstrated reuse rate and the binding compliance envelope in 2026?
50–70% cooling-tower reuse is held by a high-recovery RO at 70–85% recovery, with the lower-TDS blowdown fraction diluting the fab reject's fluoride and silica load (HydropureWater field data, 2026). IED 2.0 BAT-AELs bind regardless of discharge route, so the train must satisfy them on the sewer envelope and on any surface-water envelope simultaneously; UWWTD 91/271/EEC plus the 2024 amendment covers indirect-discharge pretreatment thresholds. For fluoride chemistry, the spec for HF wastewater treatment by chemical precipitation is the relevant 2026 reference. Buyers should confirm both the concentrate TDS ceiling in the FCM/OKF permit and the IED 2.0 BAT-AEL numeric values applicable to surface treatment at the submission date.
