The Desalination Trap

Why east coast capex is building the wrong solution — Australia has already spent over $11 billion on coastal desalination plants, has $5 billion more committed, and is locking in $1–2 per kilolitre operating costs forever. Meanwhile the MMA aqueduct delivers 30,000+ GL per year at near-zero ongoing energy cost. The economics are not close. Neither is the environmental case.

1. The Trap

Every Australian east-coast capital city is making the same decision in sequence. A drought hits. Dam levels fall. Public anxiety rises. The government commissions a desalination plant. The plant gets built at a cost between one and six billion dollars. The drought breaks. The plant is mothballed but still costs money to maintain. The next drought arrives. The plant is restarted. Repeat. After two decades of this cycle, the plant is expanded at a further cost in the billions because population growth has outpaced what the original plant could deliver.

This is what has happened in Sydney (Kurnell, 2010, $1.8 billion, mothballed 2012, restarted 2019, now doubling). It is what happened in Melbourne (Wonthaggi, 2012, $5.7 billion, idle most years, expansion under consideration). It is what happened in Adelaide (Port Stanvac, 2011, $2.2 billion, largely idle). It is what happened in Brisbane / Gold Coast (Tugun, 2009, $1.2 billion, drought-only operation). It is what is happening in Perth, which is the exception only in that it actually runs its plants at capacity because Western Australian rainfall has structurally declined.

The cycle is now repeating. Sydney is in detailed planning for a $1.5–2 billion doubling of Kurnell to 500 ML/day, with construction 2027 to 2030. Melbourne is in early-stage planning for further expansion of Wonthaggi. South-east Queensland is considering a new plant at Stradbroke Island. Adelaide is exploring options. Every east-coast capital has the same project on the desk: another large desalination plant or expansion of an existing one. The combined forward capex commitment is approaching $5 billion, on top of the $11 billion already spent.

Every one of these projects locks in $1–2 per kilolitre operating costs for the life of the plant — typically thirty to fifty years. The capex is a one-time cost. The opex is forever. And the alternative — an aqueduct that captures water that is currently lost to the ocean and delivers it to where it is needed via the corridor that is being built anyway — is not on any state government’s procurement plan.

2. What Has Already Been Built

Two observations from this table are important. First, the total nameplate output of every Australian desalination plant combined is approximately 540 GL per year. This is less than two percent of the 30,000 GL annual target of the MMA Sovereign Aqueduct Network described in Memo 14. The entire desalination industry, built across two decades at a cost of more than $11 billion, produces less freshwater than a single year of MMA aqueduct flow at design capacity.

Second, every one of these plants discharges its waste brine into coastal waters. Three of the six discharge into semi-enclosed bays or sounds — Botany Bay, Gulf St Vincent, and Cockburn Sound — where tidal exchange with the open ocean is limited and the salinity impact accumulates locally. This is the environmental problem the industry would prefer not to discuss. It is addressed in section 6 of this memo.

3. What Is Being Planned

The forward pipeline of Australian desalination expansion is large, advancing through procurement, and approaching final investment decisions in several states without a national-scale alternative being formally considered.

The Sydney expansion is the most immediate decision and the most fully committed. Sydney Water commenced detailed planning in 2023, procurement is scheduled to commence in 2026, contract award is targeted for the same year, and construction is to begin in 2027 with completion in 2030. This memo is being written in May 2026. The contract award is months away. Once contracted, the project is effectively locked in: cancellation costs, contractual obligations, and procurement timelines make reversal expensive and politically difficult.

The decision window for considering a national alternative to the Sydney Desal expansion is open now. It will close within twelve months. Once the doubling is contracted, Sydney will be committed to a further $1.5–2 billion of capex and approximately $50–100 million per year of additional operating cost for the next thirty to fifty years — before the doubled brine load on Botany Bay is even considered.

4. The 30-Year Demand Wall

The desalination expansion programme is not a hypothetical. It exists because east coast water demand is rising and the planners running these utilities know it. Understanding the demand curve they are responding to is essential to understanding the choice. Australia’s east coast will need substantially more freshwater over the next thirty years than it consumes today. That water must come from somewhere. The question is where.

4.1 Population Growth

Greater Sydney is projected to grow from approximately 5.4 million today to approximately 8 million by 2056 — an increase of around 2.6 million people. Greater Melbourne is on a similar trajectory, projected to reach 8–9 million by 2056. South-east Queensland will exceed 5.7 million. Across the three major east coast metropolitan regions, an additional 6–7 million Australians will be drinking, cooking, bathing, gardening, and flushing within the next thirty years. At an average residential consumption of approximately 200 litres per person per day, that is somewhere in the order of 400–500 gigalitres of new annual residential water demand from population growth alone — before any increase in commercial, industrial, or agricultural consumption is counted.

4.2 Industrial and Commercial Demand

Population growth is only one driver. The transition of Australian industry toward electrified manufacturing, data centre development, green hydrogen production, and large-scale agriculture for export markets all require water at industrial scale. A single gigawatt-scale data centre consumes 1–3 gigalitres of water per year for evaporative cooling. Green hydrogen production via electrolysis requires approximately 9 litres of high-purity water per kilogram of hydrogen — meaning a 1 Mt per year hydrogen export industry consumes 9 GL of water per year. Australian beef, dairy, and horticulture export industries are water-intensive and growing. Climate-controlled vertical farms, advanced food processing, biotech manufacturing, and battery production all add to industrial water demand.

None of this industrial demand exists in current east coast water utility forecasts at credible scale. The forecasts assume a continuation of residential and small-commercial demand patterns. The actual demand trajectory, when industrial decarbonisation and AI infrastructure are properly counted, is materially higher than what current desalination plans are sized to meet.

4.3 Climate-Driven Rainfall Decline

The supply side is also moving in the wrong direction. South-eastern Australia has experienced a long-term decline in cool-season rainfall since the 1990s, attributed to a southward shift in the rain-bearing westerlies driven by climate change. The CSIRO and Bureau of Meteorology State of the Climate reports document a 10–20 per cent reduction in cool-season rainfall across the southeast over recent decades, with further decline projected. Warragamba Dam — Sydney’s primary water source — experienced its lowest-ever inflows during the millennium drought and again during 2017–19. The Murray-Darling Basin’s long-term inflow trend is downward. Perth, on the same dynamic at greater intensity, has already lost roughly two-thirds of its historical rainfall-fed runoff.

The implication is that east coast water utilities are simultaneously facing rising demand and falling rainfall-dependent supply. The gap between what the dams will reliably deliver and what the cities will need is what the desalination expansion programme is sized to fill. The expansion is not optional in any of the published water strategies — the only question is what fills the gap.

4.4 The Volume That Must Come From Somewhere

Combining population growth, industrial demand, and rainfall decline, the east coast water gap over the next thirty years is in the order of 1,000–2,000 gigalitres per year of new firm freshwater supply — roughly two to four times the entire current Australian desalination industry’s nameplate capacity. This is the volume that must come from somewhere. There are only three plausible sources at this scale:

• Manufactured freshwater from seawater — continued and expanded desalination, with its associated electricity demand, chemical inputs, membrane consumption, and hypersaline brine discharge into coastal waters.
• Recycled water — treated wastewater reused for industrial or potable supply. Modest contribution at the scale required, public acceptance remains limited for potable reuse, energy-intensive treatment is similar in scale to desalination.
• Northern monsoon and inland flood capture — the MMA Sovereign Aqueduct Network, capturing water that currently flows to the ocean and delivering it via the MMC corridor to where it is needed.

The first two options are extensions of the existing approach. The third is fundamentally different. It does not manufacture water by separating salt from seawater. It does not energy-intensively process water that has already been consumed. It captures water that nature delivers in abundance every wet season and that Australia currently loses entirely. The thirty-year demand wall does not change because of this memo. Where the water comes from to meet it can.

5. The OPEX That Never Stops

Desalination plants are not built and then forgotten. They are industrial facilities that consume large quantities of energy, chemicals, and consumables continuously. Operating costs in the desalination industry are well documented internationally and account for fifty to seventy per cent of the total lifecycle cost of every cubic metre of water produced.

5.1 The Energy Bill

Reverse osmosis desalination requires high-pressure pumps to force seawater through semi-permeable membranes against the osmotic pressure of dissolved salt. The energy required is approximately 3–4 kWh per cubic metre of freshwater produced, including pre-treatment and post-treatment. Sydney Desal at full 250 ML/day output consumes approximately 38 megawatts continuously — equivalent to about 30,000 average homes. The doubled plant will consume approximately 76 MW continuously.

Energy is the largest single component of desalination operating cost — typically 35–45 per cent of total OPEX. At Australian commercial electricity prices of 10–20¢/kWh, energy alone adds $0.30–$0.80 per cubic metre of freshwater. Across the doubled Sydney plant at 91 GL per year, that is $27–73 million of electricity every single year — in perpetuity, scaling with electricity prices.

This is grid electricity. It is generated from coal, gas, and renewables across the eastern Australian grid. It is not free. It is not renewable by default. Sydney Desal’s claim of one hundred per cent renewable energy is a market-based instrument — the plant draws power from the grid like any other industrial consumer and the renewable claim is satisfied through certificate purchases. The grid load is real regardless of how the accounting is structured.

5.2 Membranes — Replaced Every Three to Five Years

The semi-permeable membranes that perform the actual desalination work foul, scale, and degrade under continuous operation. Standard reverse osmosis membranes have a working life of three to five years before they must be replaced. A large desalination plant contains tens of thousands of individual membrane modules. Membrane replacement is one of the largest recurring expenses in the plant’s budget — accounting for 5–15 per cent of total OPEX and representing a periodic capital expenditure of tens of millions of dollars every replacement cycle.

Sydney Desal contains approximately 36,000 individual membrane elements. At a replacement cost of approximately AUD 800–1,200 per membrane element installed, full replacement is in the range of $30–45 million every three to five years. The doubled plant scales these costs proportionally.

5.3 Pre-Treatment and Filtration

Before seawater reaches the reverse osmosis membranes it must be pre-filtered to remove suspended solids, plankton, oil, and biological contaminants. Pre-treatment filtration uses multi-media sand filters, cartridge filters, and ultrafiltration membranes. These filters require continuous backwashing with seawater and periodic chemical cleaning. Cartridge filters require replacement on a continuous schedule — thousands of units per year for a plant the size of Sydney Desal. Multi-media filter sand must be periodically replaced when fouling becomes irreversible. The combined cost of filter consumables and pre-treatment chemicals is typically 5–10 per cent of plant OPEX — another $5–15 million per year for a 250 ML/day plant.

5.4 Chemical Inputs

Modern reverse osmosis desalination uses substantial quantities of process chemicals throughout the treatment train:

• Antiscalants — injected into the feed water to prevent mineral scale formation on the membranes. Typical dosing 2–5 mg per litre of feed water. Cost $2–5 per kilogram.
• Coagulants — ferric chloride or polyaluminium chloride used in pre-treatment to remove suspended solids. Cost $0.30–0.80 per kilogram.
• Acid — sulphuric or hydrochloric acid used for pH adjustment to prevent scale formation. Cost $0.10–0.30 per kilogram.
• Sodium bisulphite — dechlorination chemical applied before the RO membranes to protect the membrane material from chlorine attack.
• CIP cleaning chemicals — sodium hydroxide, citric acid, and surfactants used in clean-in-place membrane cleaning cycles every few months.
• Post-treatment minerals — calcium hydroxide and carbon dioxide for remineralisation, chlorine or chloramine for disinfection, fluoride for dental health compliance.

The combined chemical bill represents 5–15 per cent of plant OPEX — another $5–20 million per year depending on plant scale and feed water quality.

5.5 Labour, Maintenance, and Insurance

A large desalination plant employs 50–100 staff in continuous operation across shift work, plant management, engineering, and maintenance. The Sydney plant employs approximately 70 staff. Annual labour cost in Australian conditions is in the range of $10–15 million per year. Equipment maintenance — pumps, valves, instrumentation, the energy recovery devices that are critical to plant economics — adds a further 10–15 per cent of OPEX. Insurance, regulatory compliance, and administrative overhead add the balance.

5.6 Total OPEX in Australian Conditions

This is the running cost of Sydney’s doubled desalination plant alone — a single facility serving one city. Scaled across the five existing Australian desalination plants and the proposed expansions, total national desalination OPEX at full operation runs to approximately $500–800 million per year, every year, in perpetuity. The capex is the one-off cost. This is the permanent cost.

6. Two Different Approaches — The Comparison That Matters

The desalination figures in this memo are real and documented. They come from publicly disclosed operator data, IPART pricing determinations, and international cost benchmarks for reverse-osmosis seawater desalination. They are not contested.

The MMA Sovereign Aqueduct Network is at an earlier stage of cost development. The detailed capital cost is being scoped by the Sovereign Build Corporation as part of the broader MMC corridor programme and is not finalised at the time of this memo. What can be stated with confidence is the structural logic of the two approaches — the physical inputs each requires and the ongoing operating burden each carries.

The fundamental driver is physics. Desalination manufactures freshwater by pushing seawater through membranes against the osmotic pressure of dissolved salt. The thermodynamic cost of separating salt from water at scale is enormous and permanent — that is what the energy bill, the membrane replacement, the chemical dosing, and the brine discharge are all consequences of. The aqueduct moves water that is already fresh. The thermodynamic cost of moving water is comparatively trivial, especially when the pumping energy is curtailed solar that would otherwise be wasted.

The MMA approach has costs — it requires the corridor to be built, it requires the Alice Hub storage to receive the water, and it requires the elevated viaduct feeder network described in Memo 14. Those costs are real and they are being scoped properly. The argument this memo makes is not that the MMA approach is free. It is that, once the corridor is built for its other purposes, the marginal cost of delivering water through it is structurally different from the marginal cost of manufacturing water by reverse osmosis — in every category and across every kilolitre delivered for the life of the infrastructure. → See: Alice Hub — MMA Memo 5. The Sovereign Aqueduct Network — MMA Memo 14.

7. The Product Water Question

Brine is what leaves the back end of the desalination plant. The other question is what comes out of the front. Reverse osmosis does not simply filter seawater — it strips it to demineralised purity and then engineers a drinking water product through a sequence of chemical additions. The result is delivered through the same taps as conventional dam-and-treatment water, but its production chemistry is materially different. The peer-reviewed literature, including WHO technical reports, identifies several open questions about long-term consumption that are worth understanding honestly.

7.1 What Actually Gets Added in the RO Process

An operating reverse osmosis plant uses approximately a dozen chemical inputs across pre-treatment, membrane protection, and post-treatment stages. Where they end up depends on the chemical:

The honest picture: most of the heavy chemistry — coagulants, the bulk of antiscalants, the CIP cleaning chemicals — does not reach the consumer because it is either captured in pre-treatment sludge or carried out in the brine reject. That fact is part of why the brine load is a marine discharge concern rather than a drinking water concern. But residual antiscalants — phosphonate-based compounds added continuously to feed water and only partially rejected by the membrane — do appear at low concentrations in product water. The long-term health data on phosphonate antiscalants in drinking water at lifetime exposure levels is limited. The compounds are widely regarded as low-toxicity at the residual concentrations measured. They are not extensively studied at thirty- to fifty-year human exposure timescales.

7.2 The Demineralisation-Remineralisation Cycle

Reverse osmosis is designed to remove essentially everything dissolved in seawater — salt, of course, but also the calcium, magnesium, sulphate, potassium, bicarbonate, and trace minerals that occur naturally in fresh water and that humans have evolved drinking. Permeate water leaving the RO membrane has total dissolved solids close to zero. Calcium and magnesium are below detection limits.

This water cannot be delivered to consumers as it leaves the membrane. It is mildly aggressive to pipework, lacks essential dietary minerals, and tastes flat. Post-treatment remineralisation — adding calcium hydroxide and carbon dioxide to bring hardness up to a target value — is therefore standard practice at every Australian desalination plant. The WHO has acknowledged demineralised water as a health consideration since the 1980s, with long-term consumption of insufficiently remineralised desalinated water linked in some populations to electrolyte disturbances, cardiovascular risk markers, and reduced bone mineral density.

Australian plants do perform remineralisation. The question is not whether minerals are added back — they are. The question is whether industrial post-treatment delivers the same biological outcomes as drinking naturally mineralised water at lifetime exposure. The published evidence here is genuinely mixed. The remineralisation step is also another continuous chemical input adding to the operating cost discussed in Section 5.

7.3 Boron — The Chemistry RO Does Not Solve Cleanly

Seawater contains approximately 4–5 mg/L of boron. The boron concentrations matter because boron compounds are linked to reproductive and developmental effects in animal studies at elevated lifetime exposure. Standard high-rejection seawater RO membranes remove approximately 73–90% of boron in a single pass — leaving 0.5–1.5 mg/L in the permeate stream depending on temperature, pH, and membrane condition.

The WHO guideline for boron in drinking water is 0.5 mg/L. Achieving this from RO seawater desalination requires additional treatment — typically a second RO pass at elevated pH, specialised high-boron-rejection membranes, or selective ion exchange. All of these add capital cost, operating cost, and energy consumption to the plant.

The Australian Drinking Water Guidelines (ADWG) set the boron limit at 4 mg/L — substantially more permissive than the WHO 0.5 mg/L recommendation. Australian desalination plants typically comply with the ADWG limit but the standard itself is more lenient than the international guideline. This is a genuine open question worth flagging: Australian consumers of desalinated water may be exposed to boron concentrations higher than the WHO considers desirable, within the domestic regulatory standard. The peer-reviewed literature on long-term health implications at the Australian regulatory ceiling is limited.

7.4 What the Aqueduct Water Does Not Need

Freshwater captured from northern monsoon flow and Lake Eyre Basin floods is, biologically speaking, the same category of water that fills Warragamba, Thomson, and Hinze dams — naturally mineralised surface water from rainfall and runoff. It has a mineral profile broadly compatible with what humans have evolved drinking. It contains the dissolved calcium, magnesium, bicarbonate, and trace elements that participate in normal mammalian physiology.

The aqueduct water still requires treatment before consumption — standard sedimentation, filtration, disinfection, and fluoridation, the same processes applied to conventional Australian dam supplies. What it does not require is demineralisation followed by industrial remineralisation. It does not require boron removal. It does not require continuous antiscalant dosing. The chemical inventory of the treatment train is substantially smaller, the chemistry is well-characterised by a century of municipal water supply experience, and the residual chemical exposures are inherently lower. This is not the central argument for the MMA aqueduct. The capex and opex case made in earlier sections is. But the product water question is one more dimension where the two approaches differ in the favour of the aqueduct.

8. The Brine Question — Honest About What Is Known

Hypersaline brine discharge is the dimension of desalination most often discussed in environmental terms. It deserves honest treatment. The evidence base is global, the picture is genuinely mixed, and Australian site-specific monitoring has — to date — not detected the catastrophic impacts seen at some overseas locations. The risk is real but it is context-dependent, and the most important issues are about scale over time rather than any single plant’s current performance.

8.1 What the Brine Actually Is

Brine from a reverse osmosis plant is not just salty water. It typically contains:

• Salinity at approximately 60–70 g/L — roughly twice the salinity of ambient seawater (~35 g/L). The brine is denser than the surrounding water and tends to sink rather than mix where currents are weak.
• Residual antiscalants, coagulants, and acidified water from pre-treatment and membrane protection chemistry — diluted, but present in continuous discharge.
• Trace metals — copper, iron, and others released from the high-pressure pumping equipment as it corrodes under saline service.
• Backwash sludges from the pre-treatment filters — periodically discharging sediment, biological material, and concentrated chemical residues.
• Elevated temperature — typically 1–3°C warmer than ambient seawater due to friction heating in the pump and recovery systems.

8.2 What the Global Evidence Shows

Two decades of peer-reviewed research from desalination operations around the world establish a clear pattern: impact depends overwhelmingly on the receiving environment. Discharge into well-flushed open coastal waters with strong currents typically produces small, localised effects measured in tens to perhaps a hundred metres of seabed alteration around the diffuser. Discharge into poorly flushed environments — semi-enclosed bays, shallow sounds, gulfs with weak tidal exchange — produces materially different results: brine plumes that creep along the seabed for kilometres, benthic community shifts, seagrass meadow contraction, and accumulation of treatment chemistry that local tidal flushing cannot remove on the timescales the discharge continues.

The documented evidence comes primarily from international sites with poor flushing characteristics:

• Alicante, Spain — Posidonia oceanica seagrass meadows. Studies have demonstrated that Mediterranean seagrass meadows experience meadow contraction and physiological stress at salinity elevations of only 1–2 ppt above ambient sustained over months. Posidonia is a foundation species and slow to recover.
• Persian Gulf — macrobenthic community changes. The combination of poor flushing, multiple co-located plants, and elevated baseline salinity has produced documented changes in benthic invertebrate community composition near several Gulf desalination outfalls.
• Israel and California — benthic effects within designated mixing zones. Sustained salinity elevation in the immediate discharge zone produces measurable shifts in soft-sediment invertebrate communities and, in some cases, sessile species mortality.

The consistent finding across the literature is that brine impacts in well-flushed environments tend to be on the scale of tens of metres, while impacts in poorly-flushed environments can extend kilometres and persist while discharge continues.

8.3 What Australian Monitoring Has Found So Far

Australian desalination plants have been engineered with the global evidence in mind. Sydney Desal uses high-velocity diffuser nozzles at its offshore outfall in the open Tasman Sea east of Kurnell — not in Botany Bay itself — specifically to maximise brine dilution. A six-year independent monitoring programme led by the University of New South Wales (with NSW Fisheries Research and Southern Cross University), conducted across the construction, operating, and idle phases of the plant, found no significant ecological impacts at the broader site level — with elevated salinity dissipating to near background within approximately 100 metres of the discharge point. The Gold Coast plant has produced similar findings at its open Pacific Ocean outfall.

Cockburn Sound is the site where the picture is less reassuring. Cockburn Sound is a shallow, partly enclosed body of water with restricted exchange to the open Indian Ocean and significant pre-existing industrial pressure. The Kwinana desalination plant discharges into this environment. The receiving environment is exactly the type the global evidence identifies as most vulnerable to long-term cumulative impact. Independent research has flagged Cockburn Sound’s declining ecological condition over recent decades, attributed to a combination of industrial discharge, port operations, urban runoff, and the desalination plant’s contribution. Disentangling individual contributions is genuinely difficult — but the receiving environment is the wrong one for hypersaline discharge by every dimension of the global evidence base, and the cumulative trend is downward.

The honest summary is this: where Australian plants discharge into well-flushed open coastal waters with engineered high-velocity diffusers, monitoring has not detected significant ecological impact at the spatial scales measured. Where plants discharge into shallow, semi-enclosed receiving waters, the picture is more concerning and the long-term trend deserves scrutiny.

8.4 The Cumulative and Long-Term Question

The current monitoring framework is good at answering one question well: is this specific plant’s discharge producing measurable impact at this specific compliance point at this specific time? It is less good at answering the questions that actually matter for the next thirty years of east coast water policy:

• What is the cumulative impact across all Australian plants combined, monitored at the national coastal-environment scale rather than the per-plant compliance scale? Six hundred megalitres per day of hypersaline brine entering Australian coastal waters today, doubling by the early 2030s, is a continuous national-scale pollutant load that has never been formally evaluated as a single phenomenon.
• What is the multi-decade impact? The Sydney monitoring programme ran for six years. The plants are designed for thirty- to fifty-year operating lives. Long-term cumulative effects on slow-responding ecosystems — particularly seagrass meadows, sessile communities, and long-lived benthic species — are not visible in six-year studies.
• What happens at doubled capacity? Sydney Desal’s monitoring was conducted at 250 ML/day output. The expansion will take the plant to 500 ML/day. Brine load doubles. Diffuser performance at twice the throughput is an engineering question that has been modelled but not empirically demonstrated at scale. Whether the “impact within 100 m” finding extends to a doubled discharge is genuinely an open question.
• What are the synergistic interactions with climate change and other stressors? Australian coastal waters are warming, acidifying, and experiencing increased extreme weather. Brine discharge does not occur in a stable baseline — it occurs as one stressor among many in coastal ecosystems already under pressure.

None of these questions have settled answers. The precautionary principle would suggest that committing another $5 billion of capex to expand discharge volumes while these questions remain open is worth pausing on — particularly when an alternative architecture exists that produces no marine discharge at all. The brine question is not the largest argument in this memo. The capex and opex case is. But the brine dimension is one more reason to weigh the alternative carefully before contracting the next round of expansion.

9. The Stranded Asset Problem

Once the MMA aqueduct is operational and delivering water to AI campuses, hydrogen production, corridor towns, and southern irrigation at a fraction of desalinated water’s cost, every existing east coast desalination plant becomes a stranded asset. The political and financial reality of stranded infrastructure assets is well understood. The asset still exists. It still requires maintenance. It still costs money to keep operational or to mothball. The contracts that fund it through capacity payments to private operators continue regardless of whether the plant runs. The local communities that host it continue to bear the environmental burden of its discharge whether or not the city it serves needs its water.

Melbourne’s Victorian Desalination Plant is the test case. Built at $5.7 billion in 2012, the plant has produced only 505 GL total over nine years — an average annual utilisation of approximately 38 per cent of nameplate capacity. Victorian water consumers pay capacity charges to the operator every year regardless of whether water is ordered. The plant exists primarily as drought insurance. When the MMA aqueduct delivers continuous freshwater to Victoria via the corridor extension, the Wonthaggi plant’s capacity payments continue under the existing contract until the contract expires — while the water it produces is no longer needed and cannot compete on price.

This is the trap. Each new desalination plant locks in additional stranded asset risk that grows with every commitment to a national alternative. The Sydney expansion will be procured in 2026. If contracted, Sydney is locked into the doubled plant’s capacity charges for thirty to fifty years. If the MMA aqueduct is operational before that contract expires, Sydney pays both — the desal plant for being there, and the aqueduct for the water that is actually being delivered. This is not hypothetical. This is the structure of every existing infrastructure capacity payment contract in the water sector.

The window to avoid this stranded asset compounding is now. The decision being made in 2026 on the Sydney doubling determines whether one east coast city is locked in for another generation of $50–100 million annual capacity payments alongside whatever the MMA aqueduct ultimately delivers. The decisions in Melbourne, SEQ, Adelaide, and Perth follow in 2027–2030 with the same dynamic.

10. The Choice

The east coast water security problem is real. The droughts will come. The dams will fall. The population is growing. Every state government has a legitimate obligation to plan for rainfall-independent water supply. The desalination industry has supplied an answer to that obligation that is technically functional — reverse osmosis works, the plants produce drinking water, and the system is reliable when running.

The MMA Sovereign Aqueduct Network offers a different answer to the same obligation. It does not require thermodynamic separation of salt from water. It does not require sustained gigawatt-scale grid electricity. It does not require continuous chemical inputs, membrane replacement, or any marine discharge at all. It delivers thirty to one hundred times more freshwater than the entire Australian desalination industry combined, at a fraction of the operating cost per kilolitre. It serves multiple end uses from the same infrastructure investment — AI cooling, hydrogen, towns, irrigation, and southern dam supplementation — rather than a single drinking water output.

This memo does not argue that existing desalination plants should be torn down. They have been built, they have been paid for, and they provide a genuine drought-insurance function for their host cities. They will continue to operate within their economic lives. The question this memo addresses is what is built next. Every new desalination plant or expansion that is committed before the MMA alternative is formally on the table compounds the stranded asset problem and locks in another generation of unnecessary OPEX, unnecessary brine discharge, and unnecessary opportunity cost.

Australia is at a decision point. The next $5 billion of east coast desalination capex is in procurement. Once contracted, it is locked in for half a century. The alternative is sitting in front of the Australian government as a comprehensive engineering programme with continental scale, multi-purpose infrastructure, and operating costs that approach zero. The question is whether the decision is made on the desk it deserves — in front of federal and state water ministers comparing both options on the merits — or whether it is made by default through the procurement timeline already in motion. Time is short. The Sydney contract is twelve months away.