Why Amman is a water-stress design case in 2026
Jordan is one of the most water-scarce countries on earth, so any new fab or hyperscale data hall around Amman is being planned in 2026 with water reuse, not discharge, as the baseline. Globally the semiconductor industry consumes around 210 trillion litres of water annually, almost half of which is consumed in areas facing higher-than-average water scarcity (TNFD, Feb 2026, citing Yin & Yang 2025). A single fab can use around 14 billion litres of ultrapure water (UPW) per year, and for every unit of UPW, 1.4–1.6 units of municipal water are used (TNFD, Feb 2026, citing IDE Technologies 2024).
Data halls sit in the same risk envelope. A typical data centre can use 25 million to 770 million litres of water per year, while hyperscale facilities may exceed 2 billion litres annually (TNFD, Feb 2026, citing Ceres 2025 and Hines Research 2025). 45% of data centres globally are in river basins at high risk of water availability disruptions (TNFD, Feb 2026, citing Hajonides et al. 2025), and 40% of existing fabs and over 40% of new fabs announced since 2021 are projected to be in basins with high or extremely high risk of water stress by 2030 (TNFD, Feb 2026, citing Lepawsky 2024).
The Jordanian anchor is the As Samra wastewater treatment plant northeast of Amman. MIGA-backed expansion increased the plant's treatment capacity by 37%, and it now treats around 70% of Jordan's total wastewater, meeting the needs of 3.5 million people (MIGA). For an engineer in 2026, that single fact reframes every discharge conversation: As Samra is the country's central wastewater funnel, so any new industrial stream tied into it inherits both its pretreatment envelope and its public-profile risk. The defensible 2026 design default is high-recovery reuse, with As Samra treated as a backup, not a baseline.
Process wastewater streams a fab or data hall actually generates
A fab in the Amman area typically generates CMP slurry wastewater, HF/nitric/ammonia etch rinses, IPA-bearing rinses, TMAH photoresist developer, acid-alkaline cleaning streams and scrubber blowdown. A co-located data hall contributes cooling-tower blowdown, humidification bleed, RO reject and generator or IT-room drainage. The exact concentration profile is site-specific, so engineers should obtain their own influent characterisation before sizing any stage.
What the public case record shows is that feed variability, not any single contaminant, tends to drive the train. IDE's semiconductor case study describes "highly variable feedwater conditions, including a high fouling tendency, SDI values most of the time above 5 (and at times non-measurable), and significant variation in feedwater composition (both TDS and organic matter)" (IDE). That single statement rules out a conventional RO as the recovery workhorse without pre-treatment polishing, and it forces the design toward high-fouling-tolerant RO with frequent clean-in-operation cycles.
Ultrafiltration is the workhorse between equalisation and any downstream RO or MBR. Saltworks notes that "XtremeUF ceramic and polymeric ultrafiltration removes oils, grease, precipitated by-products, particulate, microbes, and suspended solids" (Saltworks Technologies). The data-hall subset is dominated by cooling-tower blowdown chemistry — TDS, silica and biocide residuals — which sets up the silica-aware recovery ceiling discussed later. For more on UF and MBR sizing, the 2026 MBR membrane module design guide is the natural companion reading.
| Stream group | Typical constituents | Design driver for the train |
|---|---|---|
| CMP slurry wastewater | Suspended silica/ceramic particles, metals (Cu, Ni, Co), surfactants | Solids, FOG and metal removal upstream of RO |
| HF / nitric / ammonia etch rinses | Fluoride, nitrate, ammonium | pH window, fluoride and ammonia control |
| IPA, TMAH photoresist developer | Volatile organics, COD, high pH | Separate buffering, biological treatment for the biodegradable fraction |
| Scrubber blowdown | Acidic/alkaline residuals, dissolved metals | Equalisation and pH balancing |
| Cooling-tower blowdown / humidification bleed | TDS, silica, scale/corrosion inhibitors, biocides | Silica-limited RO recovery |
| RO reject / IT-room drainage | Concentrated TDS, trace metals | Brine handling and possible ZLD |
The 2026 treatment train Amman sites are converging on
The train a 2026 Amman site converges on is six stages: equalisation, coagulation plus DAF or lamella clarification, ultrafiltration, biological treatment (MBR), high-recovery RO, and a brine-finishing step. Each stage earns its place from the chemistry above and the Jordanian discharge constraints.
Stage 1 — equalisation and pH balancing for the mixed fab stream, with separate buffering for TMAH-bearing developer wastewater to protect downstream biology. Stage 2 — coagulation and flocculation followed by a dissolved air flotation system for suspended solids, FOG and precipitated metals. DAF catalogues cover 4–300 m³/h across 13 models, which is the qualitative envelope for fab flows in the Amman capacity range. Stage 3 — a hollow-fibre ultrafiltration system as a polishing and RO-protection step, paired with a multi-media pre-RO filter on the cooling-tower loop. Stage 4 — a submerged MBR membrane bioreactor for COD/BOD reduction in the more biodegradable fractions, with submerged PVDF membranes giving <1 µm filtration and a 60% smaller footprint than conventional activated sludge.
Stage 5 is where the design pays for itself. The IDE MAXH2O Pulse Flow RO was installed to "treat cooling tower blowdown and MBR effluent, together with a variable, concentrated CMP stream", and today "the system treats ~720 GPM (~4,000 m³/day) and achieves 54% recovery (silica-limited). This represents an 18% increase over recovery levels before deployment of IDE's solution" (IDE). The same case study reports "recovery of additional 54% from the upstream treatment system brine, resulting with an overall recovery rate of 88%" (IDE). That pair of numbers — 54% on a silica-limited pass, 88% on the system as a whole — is the load-bearing data point a 2026 Amman engineer should quote to management. The RO stage should be a high-recovery industrial RO system sized against silica saturation, not against TDS, since "a pre-existing conventional RO to struggle to the point that it could not operate reliably, even with weekly CIP" (IDE) is the failure mode the design must prevent. Stage 6 — either a smaller ZLD finishing step (evaporator/crystalliser) or controlled discharge, with the choice driven by As Samra's pretreatment envelope rather than by the upstream train.
| Stage | Function | Key 2026 parameter / data point |
|---|---|---|
| 1. Equalisation / pH balance | Buffer flow and chemistry swings; isolate TMAH | Site-specific; size for peak shift flow |
| 2. DAF / lamella | Suspended solids, FOG, precipitated metals | 4–300 m³/h range across 13 DAF models |
| 3. UF (hollow-fibre) + MMF | RO protection, turbidity and particulate cut | 0.03 µm cut; feeds up to 300 NTU (qualitative) |
| 4. MBR | COD/BOD reduction on biodegradable fraction | Submerged PVDF, <1 µm, ~60% smaller footprint than CAS |
| 5. High-recovery RO (pulse-flow) | Silica-limited recovery, reuse to cooling/UPW | ~720 GPM at 54% recovery; 88% system recovery (IDE) |
| 6. Brine handling / ZLD | Concentrate disposal or solidification | Decision driven by As Samra envelope |
Reuse, discharge or ZLD — the 2026 decision for Amman
For a 2026 Amman-area site there are three realistic outcomes, and the choice between them is set by water-risk and capital-expenditure pressure rather than by effluent rules alone. Option (a) is treat-to-discharge into the As Samra catchment, which means the plant must meet As Samra's pretreatment envelope and accept the public-profile risk of sending industrial flow into a 3.5 million-person plant. Option (b) is treat-to-reuse, with RO permeate feeding cooling-tower makeup, landscape irrigation or UPW polishing pre-feed — the engineering default in 2026. Option (c) is treat-to-ZLD with brine solidification, which removes the discharge variable entirely but raises both capex and energy intensity.
The capital context pushes hard toward reuse. Data-centre capex is expected to exceed USD 1.7 trillion globally by 2030, and demand for microchips is "on track to exceed USD 1 trillion by 2030" (TNFD, Feb 2026, citing McKinsey 2025 and Deloitte 2025). That means any new Amman-area project is being designed against investor and supply-chain water-risk scrutiny, not just local effluent rules. 40% of existing fabs and over 40% of new fabs since 2021 are projected to be in basins with high or extremely high water-stress risk by 2030 (TNFD, Feb 2026, citing Lepawsky 2024), so a discharge-only design is an asset-liability risk before it is a compliance risk.
The defensible 2026 Amman stance is high-recovery RO to reuse for cooling-tower and UPW makeup, with ZLD reserved for the final brine concentrate; discharge to As Samra is the backup, not the baseline. The same logic, applied to a different Asian water-stress basin, is laid out in the 2026 semiconductor and data-hall guide for Nagoya, and the upstream DAF-vs-clarifier trade-off is examined in the DAF vs clarifier selection guide for semiconductor wastewater. The UPW-side growth assumptions that drive the reuse target are quantified in the 2026 UPW scale-up tool.
| Outcome | What it requires | Best fit in 2026 Amman |
|---|---|---|
| Treat-to-discharge to As Samra | Full pretreatment to As Samra envelope; accept public-profile risk | Backup path, not baseline |
| Treat-to-reuse (cooling/UPW makeup) | High-recovery RO (54–88%); cooling chemistry programme | Default for new fabs and data halls |
| Treat-to-ZLD | Evaporator/crystalliser; high energy and capex | Reserved for the final brine concentrate |
Compliance and operational pitfalls specific to Jordan
As Samra's pretreatment envelope sets the realistic discharge ceiling for any industrial tie-in. The qualitative checklist is the pH window, heavy metals (especially Cu, Ni and Co from CMP), fluoride from HF-bearing streams, ammonia from TMAH, and residual oxidants or biocides carried over from cooling-water treatment. Engineers should request the current As Samra industrial-discharge schedule and treat the limits as binding from day one of design.
Silica scaling is the explicit ceiling on RO recovery. The IDE case study records 54% recovery "silica-limited" (IDE), which is the design signal: treat silica, not TDS, as the pinch point when sizing reuse rates. Variable feed conditions break conventional RO, and the IDE record is explicit that "a pre-existing conventional RO to struggle to the point that it could not operate reliably, even with weekly CIP" (IDE); that failure mode justifies either a high-fouling-tolerant RO design or a parallel pulse-flow unit with frequent clean-in-operation cycles. For the cooling-water side, an automatic chemical dosing system and an on-site chlorine dioxide generator keep biocide residuals in the reuse envelope, with the dosing logic covered in the 2026 auto-dosing engineering guide. The data-hall-specific risk is that cooling-tower blowdown and humidification bleed dominate the water-side effluent, so any reuse scheme has to be coupled to the cooling-water chemistry programme to avoid re-poisoning the RO with silica, scale inhibitors or corrosion inhibitors.
Frequently Asked Questions
What does a defensible 2026 treatment train for an Amman fab or data hall look like?
Equalisation, coagulation plus DAF, ultrafiltration, an MBR, a high-recovery RO (silica-limited) and a brine-finishing step. The IDE case study records 54% recovery on a silica-limited RO pass and 88% system recovery on fab CMP, MBR and cooling-tower streams (IDE), which is the load-bearing data point to quote to management and to the regulator.
Is direct discharge to As Samra still a viable 2026 baseline?
No — As Samra treats around 70% of Jordan's national wastewater for 3.5 million people (MIGA), so a discharge-only design inherits both the pretreatment envelope and the public-profile risk. Treat it as a backup, not a baseline, and design the train for high-recovery reuse as the default.
What inputs do I need from a supplier before I can price a 2026 Amman reuse scheme?
Ask for a site-specific influent characterisation (TDS, silica, FOG, Cu/Ni/Co, fluoride, ammonia, TMAH, COD), the current As Samra industrial-discharge schedule, the design feed-temperature range, and a silica-limited RO recovery calculation rather than a generic TDS-based figure. Without those, any quoted capex is not yet defensible to management.
How do I select a treatment-train supplier for a Jordanian fab or data-hall project?
Score suppliers against three checks: documented performance on a silica-limited, high-fouling RO at the recovery range you need (the IDE record is a benchmark, not a guarantee); ability to supply the full train from DAF through MBR and RO with one process guarantee; and willingness to commit to As Samra-compatible discharge chemistry and to a Jordanian reference installation. Reject any proposal that quotes recovery against TDS alone or that treats the brine as an afterthought.
