The WaterCentury

Meet the players

Before we pull any levers, it is worth knowing who is actually holding them. There are five that we need to know.

Government makes the big rules: what gets built, what standards apply, what the law requires. Regulators set the targets, control chunks of spending and are supposed to enforce the rules. Water companies and public water bodies run the actual system - the pipes, reservoirs, treatment works and meters. Agriculture and business control most of the water used outside our homes. And then there is us shmucks - who control some of the demand but, more importantly, have the civic power everyone else answers to eventually.

One thing to establish early, because it changes answers throughout: the UK does not run one water system. England’s companies are private, overseen by a regulator called Ofwat. Wales’s biggest is a not-for-profit. Scotland’s is publicly owned. Northern Ireland’s is government owned and depends on public funding. Different ownership, different funding, different regulators, different needs - so when the answer changes at a border, I will say so.

That is all you need for now. The acronyms - Ofwat, EA, SEPA, NRW, RAPID - will introduce themselves when they actually enter the story.

Can’t we just
use less?

Yes, and this is the biggest single part of England’s plan. For better or for worse.

Start with an ordinary household. In 2024-25 the average person in England used 136.5 litres a day; the target is 110 litres by 2050. So the national strategy is asking each of us to find about 30 litres. Only?

And here is the number that makes the target look suddenly achievable: metered households average 122 litres a head. Unmetered ones average 171. Be careful with that pair, because it compares two groups of people rather than measuring what a meter does, and metered and unmetered homes differ in more than their meters. But hold the two figures next to the national ask: the gap between them is 49 litres, and the whole country is being asked to find about thirty.

Now the politics, straight away. “You’ve allowed billions of litres to leak from the network, bills are rising, and now you’re going to financially penalise me for taking a bath?” That lands, because it is reasonable. The fairness problems are real too: bigger families need more water, and a hidden leak in a metered home registers as your consumption and hands you the bill.

Saving water should mean wasting less - better appliances, meters, fixed loos - not living permanently as though it is a drought.

Which raises the obvious next question. Hang on. How much are the companies losing?

About 2.9 billion litres a day. Roughly a fifth of everything put into the pipes.

So fix the leaks, obviously. And they should. But a network is thousands of kilometres of mains and joints built over generations, and the catch is brutal in its simplicity: the obvious leaks are the cheap ones. Every litre found after that costs more than the last. The target was never zero leaks - it is the lowest leakage worth paying for, given what the alternatives cost.

The key question is, did they do it? Ofwat told the English and Welsh companies to cut leakage 16% between 2020 and 2025. They managed 9%. Four companies finished the five years leaking more than when they started. The full scorecard - who promised what, who delivered - is below, and it is worth a look: same regulator, same weather, wildly different outcomes.

Can’t we just
stop ruining the water we have?

We should. But it will not make the shortage any smaller. A cleaner river is not a fuller one. I love wild water swimming so it's not that I don't care. I don't want sewage in the water either.

So we waste less of it. That only counts for anything if what is left is fit to use, which is where this investigation stopped being about volume and started being about condition.

Everyone starts with sewage. Fair enough. It turns out building towns where rainwater and raw sewage occasionally share the same pipe has created some administrative difficulties. But quick reassurance first: this is not about sewage coming out of your tap. Drinking water is treated. Pollution matters to this investigation for three less obvious reasons: dirty source water costs more to clean; when pollution is bad enough the regulator cuts how much can be pumped out of a river at all, so there is less to go round; and a drought leaves less clean water to dilute whatever goes in.

And sewage is only part of it. A river receives the farm’s nutrients and slurry (crap), the road’s metals and tyre particles, the town’s drains, the factory’s permitted discharges and accidents. Most of that is technically fixable. The problem is cost - and who pays it. The regulator’s own scoring of how companies are doing is below, and it is NOT flattering: the sector’s worst environmental score since ratings began, with serious incidents up 60% in a year. Completely unacceptable.

Then there is the farm. And that's where we found our first genuinely annoying catch. Agriculture is essential; we need food. So the question is not “should farmers use less water?” but: who gets to use scarce water, when, and what do we get in return? Smarter irrigation exists and works. Measure soil moisture, water when crops need it, capture winter water. Let's be smarter about it people!

But there is a catch. Using less water per hectare does not necessarily mean taking less water from the river overall.

A farm might cut irrigation by 20% on each field, then use that saved water to irrigate more fields. The irrigation system is more efficient, but the farm can still take exactly the same amount of water from the river.

Economists call this the rebound effect. If the goal is protecting rivers, the number that matters is not litres used per hectare. It is how much water is actually taken from the river during the dry summer months.

What the gap is made of

Everything above says that protecting rivers will not make more water. That is true, but it is only half of the arithmetic. The other half is that giving water back to rivers is, you might be surprised to hear, the single biggest reason the shortfall exists at all.

Go and ask what the 2055 gap is actually made of. It is not mostly leaks, and it is not mostly people. Across England’s five water resources regions, the largest single component is water we are proposing to stop taking: abstraction cut back so that rivers keep enough to stay alive. That is not an optional bit of environmental generosity. In many catchments we are already taking more water than the river can sustainably lose; reducing abstraction is what stops us solving the supply problem by slowly destroying the source. In the West Country that is 62% of the region’s 2055 need. In the West 60%. In the East and the South East, 58% each. Only in the North does anything else come first.

The Environment Agency says it plainly about the South East: reduction in supply to meet environmental needs is the largest driver of public water supply need there, 1,350 million litres a day of it under the Do Nothing scenario.

So a good part of the gap is not a failure. It is a decision already taken, to leave water where it does the most good. What has not been decided is where the replacement comes from.

The best company in England

Look at the top of that table. Severn Trent is the only four-star company in the sector, and it recorded a serious pollution incident in 2024. On the regulator’s scoring it is apparently as good as English water gets.

Here is the same company’s own spill record for 2025, which it publishes as open data: 91,790 discharges from 2,002 storm overflows, totalling 198,886 hours. That is 8,287 days of discharge inside a single year, from the best-rated company in the country. Give me a break.

None of that is unlawful. Storm overflows are permitted, and a company can spill within its permits all year and still score four stars, because the rating measures different things. Both numbers are official. Both are published. They simply do not describe the same world.

This is the bit I keep running into. Nobody is hiding anything. The good news and the bad news are filed by different departments, and no one is required to read them side by side.

Can’t we just
keep more of it?

Two ways to hold onto winter rain: in the landscape, and behind a wall. Only one of them counts as supply, and it is not the one people expect.

Protecting what we have does not make any more of it. Which leaves us with water that turns up free every winter but mostly runs off again before anybody can do a thing with it.

Britain has spent centuries treating its landscape like a drain. We straightened rivers, drained peat, disconnected floodplains and compacted the soil, all to move rain out of sight quickly. Now we have flooding and drought in the same year and act surprised.

The landscape is the one I wanted to win: the cheapest, the fastest and the least ugly answer here. Healthy soil, hedgerows, peat, wetlands: they slow runoff, cut flooding, filter pollution, help water soak into the ground. All of that is genuinely valuable. The Environment Agency’s own position is unusually clear, though - nature-based interventions should not be relied on to close the big supply deficits. Water entering a field is not automatically water entering a kitchen tap.

There is a middle scale here too: catch the rain where it lands. Houses, schools, warehouses and industrial estates already have enormous collection surfaces attached to them. We call them roofs. Rainwater can be filtered, stored in a tank and used for jobs that do not need drinking-quality water, particularly toilet flushing, irrigation and cleaning.

The plumbing rules already allow this. The rainwater supply has to be kept physically separate from the drinking-water network because roof water can carry bird faeces, pathogens and whatever else has settled above your head. In the regulations that makes it a Fluid Category 5 backflow risk, the highest category meaning it's treated the same as nuclear waste. That sounds absurd until you understand what it means: not that rain has been legally declared poisonous, but that nobody wants contaminated tank water siphoning backwards into the public main. The solution is wonderfully boring. An air gap. The mains can top up the tank without the two supplies ever touching.

So this one works. The limitation is just scale and a bit of legislation. A tank beside a building can replace some mains water while the rain keeps coming and for a while afterwards. Carrying enough February rain all the way through a serious August drought requires increasingly large tanks, increasingly large sites and increasingly large sums of money. Keep following that logic and eventually you have reinvented the reservoir.

That does not make rainwater harvesting pointless. Quite the opposite. A tank can take pressure off drinking-water demand.

Which brings us to the wall.

This is the answer everybody reaches for, and heck I'll reach for it too. A reservoir does not make water. It takes a winter you cannot rely on and turns it into a summer you can, which is what makes it fundamental rather than optional - and which is also why it is the first thing anyone says when the subject comes up.

And for once we can talk about actual projects instead of an abstract noun. England is progressing with ten new reservoirs and one enlargement - its biggest storage programme in decades. Thames Water and Affinity Water are developing White Horse Reservoir in the South East. Anglian Water and Cambridge Water are developing the Fens Reservoir. South West Water and Wessex Water have Cheddar 2.

Each passes through the same delivery stages: announced → funded → planning → approved → building → operating. Only the last one contains water. Everything before it is a press release that something is happening.

Now the caveat, and it is enormous. An empty reservoir is an unusually scenic hole. Capacity is what the reservoir could hold. Dependable yield is what you can still promise after a bad run of winters. Those are very different numbers, and only one of them is a supply.

Can’t we just
move it?

This is the chapter that changed my mind. There are two movement problems and the second one is barely discussed.

Say we manage that. We have saved what we can, stopped ruining what is left, and held on to more of what falls. There is still a problem, and it is the most boring one in the whole investigation (maybe you noticed it in the map above): almost none of it is where the people are.

Every map makes this look infuriatingly obvious. Wet bit here, dry bit there, pipe, done. Even the rain is against us.

Transfers genuinely make the whole system more flexible - one region’s shortage can draw on another’s abundance. England has nine big transfer schemes in development, including the Grand Union Canal Transfer (moving water along existing canals, which is conceptually delightful), Severn to Thames, and Thames to Affinity. Northern Ireland is not waiting: NI Water’s Castor Bay upgrade is already under construction - £34m to go from 95 to 115 million litres a day, with better rerouting during outages.

This is not the first attempt

Long-distance transfer is one of Britain’s oldest ideas. The Victorians built three, and all three are still running. Liverpool and Birmingham drink Welsh water; Manchester drinks the Lake District. All three arrive through pipes finished before anyone alive today was born, and not one of them uses a pump.

Lake Vyrnwy to Liverpool: 109 kilometres, the aqueducts built between 1881 and 1892, three parallel gravity pipelines that were still delivering 230 million litres a day into Prescot as recently as 2014. Thirlmere to Manchester: 154 kilometres, first water in October 1894, and still the longest gravity-fed aqueduct in the country with no pumps anywhere on its route. The Elan Valley to Birmingham: 117 kilometres, begun in 1896, falling 52 metres across the whole distance and running on that fall alone.

Three hundred and eighty kilometres of aqueduct, and a total of zero pumps between them

Britain has built large transfers since - Kielder in the early 1980s - so this is not a story about a lost golden age. It is about what the country is currently capable of deciding. The nine schemes above are the first serious programme in decades, and every one of them is still in development.

Then ask the one question every transfer must answer: what if both places are dry at once? A pipe between two reservoirs that depend on the same rainfall pattern is not resilience. Correlated drought is the whole game.

That is the movement problem everybody knows about. Here is the one I did not.

In Part One I wrote a single sentence about this and moved on: a place can have water available at the source and still struggle if the network cannot physically move enough treated water quickly enough to where demand is rising. I did not understand how literally true that was until two people who work in water explained it to me separately.

Here is the mechanism. Your drinking water sits in a reservoir near you. It is refilled by a trunk main (a big pipe that carries lots of water from the treatment works). In the example I was given, that main could take in about 40,000 litres an hour, and normal demand drew 20-30 thousand. Comfortable.

Then a long dry spell dries up the private boreholes that farms rely on, and those farms switch onto the mains. A single cattle farm can add twenty thousand litres an hour - half the capacity of the pipe, from one customer. Oh my god.

Now the local reservoir is draining faster than the pipe can refill it, and the only thing holding the line is a fleet of tankers topping it up through the night.

The mains are sized for average use, not sustained demand. Which is the answer to the thing that makes everybody furious: you can be banned from using a hosepipe while the reservoir up the road is visibly full, because the constraint was never the reservoir. It is how fast water can reach you at six o’clock on the hottest evening of the year, when everyone wants it at once.

There is an explanation for why the network is shaped like that - one of the people I interviewed walked me through it. Companies plan across fifty to a hundred years, and the planning assumption was that people would keep concentrating into denser cities, so the network was built as arteries running into those cities. Now as more people are moving to the countryside and towns are developing, we've scattered demand back out again, into places that do not currently have the infrastructure to support it. It is as much a problem for the sewage coming out as the water going in.

I have not been able to corroborate that independently, and it is second-hand testimony rather than a measured finding, so take it as what it is: the only account anyone offered me that fits the shape of the network. If there is published work that confirms or kills it, I want to hear about it.

Capacity is only half of it. Even when somebody knows exactly where the fault is, getting at it is its own bottleneck. A drinking-water inspector described marking up a burst main and being stopped by the council because it was not deemed an emergency and needed to be planned in. Another job was blocked by a landowner objecting to digging on land the company already held bought-and-paid-for access rights over. A third involved a resident complaining to the council about the tankering that was keeping his village supplied. We have to accept that the general public will make our job more difficult.

The same defect appears at the far end of the loop. What causes a combined sewer to overflow is not the total volume of a storm but the rate within a single hour. Slow the water down for even an hour and the network gets a chance to clear it.

This is also where flooding and drought stopped being two subjects for me. Too much arriving too fast at the sewer, and not enough moving fast enough through the mains, are the same defect measured at the two ends of one loop.

Can’t we just
use it again?

Used water is the one source that keeps turning up in a drought. The hard part is public trust, not engineering. WOOHOO.

Moving water does not create any either. Which leaves one source this investigation has not counted yet, and it is the only one that keeps arriving when it stops raining. We have been throwing it away.

People keep producing wastewater during droughts. We already collect it and treat it. And other countries reuse far more of it than Britain does. The water cycle has been recycling water for several billion years. In England we seemingly give it a bad PR.

The comparison is not close:

Britain recycles about one thousandth of the wastewater it treats. Spain recycles a ninth of its. Cyprus recycles almost all of it.

The missing ingredient is a decision that the litres matter. (The figures are not perfectly comparable between countries; the caveat is in the sources below.)

Now the reaction everyone has: wait - are you asking me to drink poo water?

No actually, I'm not. Nobody is proposing putting untreated sewage in the tap. The real question is whether used water can pass through enough independent barriers - treatment, advanced treatment, monitoring, an environmental buffer, treatment again - that the result reliably meets the drinking-water standard. That can be tested. And rather than being told it is “completely safe”, the safety system it has to demonstrate is below.

But before turning wastewater all the way back into drinking water, there is a simpler question to ask.

Why does all of it need to be drinking water in the first place?

We take water from a river or aquifer, treat it until it is safe to drink, pump it through miles of pressurised pipework and deliver it to every room in a building. Then we use some of it to flush toilets, wash clothes, water plants, clean floors and cool machinery. None of those jobs necessarily needs the same standard of water we pour into a glass.

That creates another kind of reuse. Instead of treating every litre back to drinking quality, you can match the water to the job. Harvested rainwater can flush a toilet. Greywater from showers and sinks can, with the right treatment, be used again. Recycled wastewater can supply irrigation, cooling, industry without first being turned into something you would serve at dinner.

And this is where the problem gets stranger. England already knows how to build separate supplies and protect the drinking-water network from them. What is less tidy is the law around what counts as water supplied for “domestic purposes” and when that water has to meet the wholesome drinking-water standard. The Drinking Water Inspectorate has itself examined whether that framework is getting in the way of dual-pipe systems and whether water quality should instead be matched to its intended use. It's enough to drive you crazy.

We have spent decades asking how to produce more drinking water before asking how much of the water we use actually needs to be drinkable.

That does not mean running untreated greywater through the washing machine tomorrow. Separate networks cost money, mistakes matter, and retrofitting a second set of pipes through millions of existing homes could be spectacularly poor value. But new housing estates, hospitals, schools, warehouses and major developments are a completely different story. If you are laying every pipe from scratch anyway, the question becomes harder to avoid: why design the next century of buildings around using drinking water for ABSOLUTELY EVERYTHING!?

Only after that question comes the uncomfortable one everybody jumps to first: if we still need more drinking water, can we safely turn used water back into that too?

That is where London, Hampshire and Poole come in.

England has real projects to interrogate: London Water Recycling, led by Thames Water, cleared its gate three decision in July 2025. Hampshire, led by Southern Water, was chosen over a desalination plant. Poole is being assessed. Three schemes, in a country that treats billions of litres of the stuff every day.

Can’t we just
make more?

Seeding clouds, and desalting the sea. Finally! One is worth zero litres. The other is the most interesting idea in this whole investigation.

Everything up to this point has been about managing the freshwater that nature hands us: keeping it, cleaning it, moving it, using it twice. Every one of those accepts the same ceiling. What if we didn't do that?

First, because someone will ask: cloud seeding. It is real technology, used in other countries, and it needs a suitable cloud already overhead - it cannot make one. No, Andy Burnham doesn't have a weather machine. Published results run from no measurable extra rain to about 20% more, with huge uncertainty. The UK does not do it and has no plans to. So for once there is no water company to yell at, no missed target, no 2030 promise. I was more disappointed by that than I expected to be, which told me something about how much I wanted one of these to be easy. Dependable British litres from cloud seeding today: zero. Sometimes the answer is just: not yet.

Now the sea. I know some of you have been waiting for this.

Part One began by effectively discarding 96.5% of Earth’s water. Too salty, get outta here, next question. Desalination asks whether we ever had to accept that.

The technology is not exotic. Reverse osmosis pushes seawater through a filter fine enough to strain out salt: freshwater on one side, concentrated brine on the other. Proven, mature, running at scale around the world. The sea is reliable. It is everything attached to the desalination that sucks the fun out of it: energy, the intake, the brine, the pipeline, planning, the grid connection, and god the financing.

And Britain has a case study in exactly this - one that keeps our investigation honest. Southern Water once proposed a major desalination plant at Fawley. It was dropped - in favour of recycling and transfers.

Desalination existing was not enough. It had to beat the alternatives, and in this instance it lost.

It is also not a hypothetical. Ofwat’s 2025-30 determination, about £8 billion of it, is already paying to pave the way for two desalination schemes, alongside ten new reservoirs, one enlargement, nine transfer schemes and a minewater treatment plant. Britain has decided to build some. It has not decided what they are for. So the question is not “should the UK desalinate?” It is: under what conditions does desalination become the better choice? Where? Powered how? Run how often?

The first problem is energy. Reverse osmosis does not simply filter seawater. It forces it through membranes at enormous pressure. Modern seawater plants typically use around 3 to 4 kilowatt-hours of electricity for every cubic metre of freshwater produced, before you start moving that water around the country. The technology has become dramatically more efficient, but physics does not disappear. Separating salt from water takes energy.

Which means a desalination plant is partly a power station problem. Build one on expensive, carbon-intensive electricity and you have one proposition. Connect it to abundant low-carbon generation and you have another. A reservoir swaps time, storing winter rain for summer. Desalination swaps dependencies. You stop depending on rain and start depending on electricity.

Then there is the bit hidden inside the word plant. A desalination plant cannot simply sit beside the sea with a pipe hanging over the wall.

First you have to get seawater to it. That means an intake capable of pulling enormous volumes from the sea without pulling the sea itself into the machinery. Screens, pumps, pipes, marine licences, ecology surveys. Depending on the coastline, that infrastructure may have to stretch offshore before you have found suitable water and somewhere sensible to put it.

Then you have to deal with what comes out the other end. Because desalination does not make the salt disappear. It makes freshwater and a smaller stream containing much more of everything you removed from it.

That brine has to go somewhere. Dump concentrated brine into poorly flushed water and the resulting saltiness will kill the things living in it. So you need another pipe, another marine structure and enough circulation at the outfall to disperse it safely. The Environment Agency specifically requires schemes to model brine concentration, mixing, dispersion and impacts on the receiving environment.

And this matters particularly in the places England actually needs the water. Being beside the sea does not automatically make somewhere a good desalination site. Shallow coastal water, sensitive habitats, currents, the seabed and existing infrastructure all determine where the intake and outfall can go. Desalination gives you an effectively enormous source of water. It does not give you an enormous number of suitable places to build the plant.

And then you might be asking yourself doesn't this just sound like more dumping waste back into our waters? And yeah, it is.

BUT there is another answer to the brine problem, and I think Britain should take it seriously: stop assuming the brine is waste.

Desalination has already done one expensive part of mineral processing for you. It has taken an enormous volume of seawater and concentrated everything dissolved in it into a much smaller stream. Salt, magnesium, calcium, potassium and traces of more valuable elements are still there. Researchers increasingly call the attempt to recover them brine mining, or rather less sexy, brine valorisation.

Lithium gets the headlines, but you can't build the case around it. Its concentration in seawater is tiny and recovering it economically remains pretty difficult. More abundant materials such as sodium chloride and magnesium compounds are much less fantastical places to start. The point is not that every desalination plant secretly contains a mine worth billions. It is that we should probably look inside the waste stream before spending money on a pipe designed to throw the whole thing back into the water.

This of course doesn't magically solve brine disposal. Recovering minerals needs more equipment and more energy, and chasing every last useful molecule eventually becomes ridiculous. Push all the way towards zero-liquid discharge and you can spend enormous amounts of energy trying to remove the final water from an increasingly concentrated soup. But there is a huge amount of territory between dump absolutely everything and extract everything.

And after you have solved the sea, you still have not solved the water supply.

The freshwater is sitting on the coast. The people who need it probably are not. Desalinated water needs conditioning to make its chemistry suitable for drinking and the network, then storage, pumping and enough pipe capacity to move it into the existing system. Sometimes the pipeline becomes one of the largest pieces of the project.

Fawley makes this wonderfully concrete. Southern Water proposed a 75 million litre a day desalination plant using seawater from the Solent. Producing the water was only part of the scheme. It also needed the marine intake and discharge infrastructure and roughly 25 kilometres of underground pipeline to move the finished water from Fawley to Testwood, where it could connect into the wider supply network.

Then there is one final trap: how often you use it.

A desalination plant built only for the worst droughts sounds economical because you only switch it on when you need it. Unfortunately, most of the cost of building the thing does not switch off with it. The intake still exists. The pipeline still exists. The membranes, grid connection and financing still exist. Run the plant rarely and all those fixed costs are divided between far fewer litres.

Britain has already had a version of this lesson. Thames Water's Beckton desalination plant was built as a drought-resilience source, originally rated at 150 million litres a day. The Environment Agency says it has only operated on a few occasions and has since been derated to 100 million litres a day. The lesson is not that desalination does not work. It is that building an extremely expensive machine and then designing the system around barely using it is a difficult way to make cheap water.

Which is why the two English schemes now being developed matter. Anglian Water's Bacton proposal could supply more than 25 million litres a day and Mablethorpe more than 50 million. Ofwat's published major-project figures put their whole-life expenditure at roughly £2.3 billion and £2.2 billion respectively. They are not giant taps capable of solving England. They are tests of whether this entire chain can be made to work where another dependable source is increasingly difficult to find.

And that changes how I think about desalination. It probably should not be the first litre Britain reaches for. Fixing a leak is easier than manufacturing the litre again. Recycling starts with water that contains far less salt. Reservoirs can store rainfall without paying an electricity bill for every litre they release.

But eventually those options run into ceilings of their own.

A reservoir can empty. A transfer can fail because the place sending the water is dry too. Efficiency can rebound. Recycling can only recycle water already moving through the system.

The sea does not run out during a drought.

That is what desalination is really buying. Not cheap water. Independence from the weather.

And suddenly “expensive” needs a comparison. Expensive compared with fixing a leak? For sure. Expensive compared with a reservoir that can be filled cheaply every winter? Yeah, probably. Expensive compared with telling farms they cannot irrigate, refusing housing connections, abstracting a river below the level nature can tolerate or discovering in August that every cheaper option was built with no margin?

Desalination is expensive right up until the alternative is not having enough water.

So the question is no longer whether reverse osmosis works. We solved that decades ago. The question is whether Britain can make the whole chain work, from intake to inland tap, at a price worth paying.

That is a much harder question.

But I think it's also finally the right one.

So what did we actually find?

At this point I was getting annoyed. Not because these ideas failed - because almost all of them worked. Just never quite in the way I wanted.

Every answer was some version of yes, but. And eventually the pattern stops being annoying and becomes the finding.

We were looking for one solution. But these things are solving completely different jobs.

A healthy river, a litre never leaked and a litre made from seawater are all valuable - for completely different reasons. Sort the ideas by the job they actually do and the whole chapter reorganises itself:

Six jobs: protect, save, store, connect, reuse, create. Not interchangeable, not in competition, and not a taxonomy anyone decided in advance. It is just where the evidence keeps landing.

Each is answering a different question, and the plan needs all six.

Protect is where you start, not because it makes water - it makes none - but because everything downstream of it gets cheaper, and a system that keeps poisoning its own sources is manufacturing a supply problem rather than solving one.

Save, which is leakage and metering and efficiency, is the only one of the six that is genuinely drought-proof: a litre never lost is a litre you still have in August.

Store turns a wet winter into a usable summer, except that landscape retention sits in the ground where nobody can dispatch it and a reservoir takes twenty odd years, so both are storage and only one of them counts as supply.

Connect is two jobs wearing one coat: moving water between regions, and having enough pipe to deliver it at six o’clock on the worst evening of the year.

Reuse keeps producing when it stops raining, because we keep producing it, and it turns up next to the cities that need it. Create is the only one that makes water which never fell on Britain. Today that means desalination and nothing else. Though Nobel Prize winner Omar Yaghi, who is extracting drinking water directly from dry desert air, might have something to say about that.

Look at that list and the shape of the problem changes. I had been reading it as a league table, trying to find the winner. There is no winner, because they are not in the same competition.

Asking whether a wetland beats a desalination plant is like asking whether a roof beats a fridge.

Which is, I think, why the public argument about water goes round in circles. Everyone is defending a different job and assuming the others are rivals. Nobody is short of good answers. We are short of a way of talking about all six at once.

And there is one more thing in that sorted board, which took me longer to see than it should have. Six jobs, and the plan does not spread its weight evenly across them.

England’s plan leans hardest on Save - over 65% of the 4.37 billion litre gap - which happens to be the job with the worst delivery record of the lot. Damn it. Hold that thought for one chapter.

Let's look back at the players

Everyone in the system has a job. There is even a law about it. Nobody has the job of checking it all adds up to enough water.

Remember the five players from the start? Now you know what all the levers actually do, the question gets sharper: who controls each job, what did they promise, and did they deliver?

The record is not a story of nothing happening. Plenty happened.

What happened was consistently smaller and later than what was promised.

The sector promised 16% less leakage and delivered 9%. Serious pollution incidents rose 60% in a year. The environmental score is the worst since scoring began. And the reservoirs and transfers meant to carry the supply side have mostly not been built yet. Not one of those failures is mysterious - every one has a name, a target, a budget and a date attached. The scorecards are back up in chapters 02 and 03 if you want to check any of it.

So who answers for that? It is the obvious question, and it does have an answer. The answer turned out to be the most revealing thing.

Here is the cast, as it stands in August 2026. Andy Burnham has been Prime Minister since July 2026. Angela Eagle has been Environment Secretary since July 2026, the third in under a year, after Steve Reed and Emma Reynolds. Ofwat, which set the leakage targets and issued the fines, is run by an interim chief executive, Chris Walters, in post since August 2025 - and it turns out Ofwat is being abolished. The Environment Agency, which handed out that worst-ever score, is led by chief executive Philip Duffy, and its chair, Alan Lovell, steps down in December 2026 with no successor named, and its water functions are being folded into a merged regulator that has no name and no leadership yet. Of everyone on that chart, Emma Hardy, who holds the water brief under Eagle, is one of the few who has worked on water continuously since 2024. Which if we're being honest, isn't that long.

Now put the dates next to each other. The leakage commitments were made for 2020 to 2025. The 60% rise in serious pollution incidents and the worst environmental score since 2011 are both the 2024 year, published in October 2025. Of the five English posts named on that chart, two were in place during the year being scored.

Three arrived after it.

That is not a defence of anybody. It is the finding. The people you would write to about the last five years are, in the main, not the people who were there for them - and by the time anyone is held to the 2050 numbers, this lot will have moved on too. The structure turns over faster than the thing it is supposed to be accountable for.

Which is why the three structural answers below matter more than any individual name in them.

The timescales do not match. A reservoir takes fifteen to twenty years. A price control runs five. A political term runs four or five. Nobody who commits to a 2050 outcome will be in post to answer for it.

The penalty lands on the wrong balance sheet. When a company is fined, customers and creditors eventually pay by some route. Thames Water’s record £122.7m is on a payment plan running to 2030. How do you make failure expensive for the people who decided, rather than the people who are billed?

And there is a duty. It is just not a duty over the total. This is the part I got wrong for weeks. I assumed nobody had been given the job in law. Somebody has. Since 2014 the Water Industry Act has placed a resilience objective on Ofwat and the Secretary of State, and it says almost exactly what you would want it to say: to secure the long-term resilience of water companies’ supply systems against environmental pressures, population growth and changes in how we use water, and to secure that companies take steps to meet, in the long term, the need for the supply of water.

So the duty exists. It is written down, in plain English, and it has been law for over a decade. You have just read what the sector did with it.

Read it closely, though, and notice what it is a duty about. It runs company by company, over each undertaker’s own supply system. There is no equivalent duty over the sum amount. Nobody is required to look at all of it together and say: this does not add up, and here is what changes.

Which is a stranger failure than an absent law, and a harder one to fix.

You cannot legislate for it twice.

So should we just nationalise it?

I have been putting this one off, because it is the first thing almost everyone says about English water and I did not want to write a pamphlet. But an investigation called Can’t we just…? cannot skip the most common answer of the lot.

The case for it is real. Water companies are regional monopolies, so you cannot take your custom elsewhere, and the only ways to make more money are to charge more or spend less. Several borrowed heavily and paid out to shareholders while the assets aged. None of that is imaginary and none of it should be waved away.

Here is the problem. Go back through the scorecards in this article and sort them by ownership instead of by company, because we are running four different models side by side.

Scottish Water is publicly owned, has no shareholders, and met its regulatory targets. It also loses about a quarter of the water it puts into supply, against roughly a fifth across England and Wales. Dŵr Cymru is a not-for-profit with no shareholders at all, and it went fifteen per cent the wrong way on leakage, the worst in the sector, having been caught misreporting its figures for years. NI Water is government owned and beat its target, its lowest leakage on record. And in England, under private ownership and one rulebook, the results scatter from companies that beat their commitments to companies that ended the period leaking more than they started.

Public ownership did not predict good performance. Private ownership did not predict bad performance. Four ownership models, and the results do not sort by ownership. Whatever explains the difference, it is not who holds the shares.

And the bill does not disappear either. Whoever owns the pipes, somebody pays for them: customers through bills, or taxpayers through government, and mostly those are the same people with a different envelope.

So the honest position is that ownership reform is an argument about incentives, transparency and who carries the risk, and it is worth having on those terms. It is not a source of water. A nationalised Thames Water would wake up on its first morning with the same leaking mains, the same undersized trunk mains, the same planning system and the same twenty-year reservoir timelines. And the resilience duty above would apply to it unchanged, because that duty does not care who owns the shares.

One more test, and it is the bluntest one

If nobody owns the outcome, you would expect that to show up in what the government actually writes down. So I ran the words through its own water white paper, published in January 2026. Seventeen thousand words.

Potable: zero mentions. Aquifer: zero. Recharge: zero. Desalination: zero. Drought: zero. Storage: zero.

Chalk streams: six, warmly, with a commitment to recover them.

Regulator: one hundred and twenty-six.

I am glad about the chalk streams. But read the two lists together. A hundred and twenty-six mentions of the regulator, and not one of the words above it. I've mentioned desalination 17 times in this article so far (now 18). Reservoirs and reuse get five each, which is five more than nothing and fewer than you would expect of the document that is supposed to be the answer. This is a paper about who is in charge of water, far more than it is about where the water comes from.

Closing the gap isn’t enough

Balancing the books is not the same as having enough. The deficit is a warning line - it was never supposed to be the ambition.

The official question is whether supply and demand can be made to balance. But a system that only balances if every project arrives on time, every efficiency target lands, the rain behaves, and nothing major breaks is not resilient - it's a slot machine. Read that sentence again knowing the sector just missed its leakage target and posted its worst-ever environmental scores.

So we need three words where most discussions use one:

Balance is enough water if everything broadly goes to plan. It is the current target, and it is worth being clear about what a target like that actually assumes: every efficiency programme landing, every reservoir opening on schedule, every transfer consented, the rain behaving roughly as modelled, and no treatment works failing at the wrong moment. Each of those is plausible. All of them at once, for thirty years, is kinda stupid.

Resilience is enough spare to absorb the things that will actually happen. A drought worse than the one that was modelled. A scheme running five years late, which on the evidence of the last five years is closer to the norm than the exception. A works down in August. Resilience is what turns a bad summer into an inconvenience rather than an emergency, and it is the difference between a system that survives and a system that copes.

Abundance is the one nobody is seemingly allowed to say out loud, so it is worth defining carefully before anyone reaches for the word extravagant. It means enough dependable water that a dry summer stops being a competition between households, food, rivers, housing and growth.

Concretely, it is the difference between restoring a river and negotiating over its remains. Between irrigating through a dry August and watching the river pay for it. Between building houses and having the water objection decide where the country is allowed to grow. Between keeping a city green as it heats, which is a public health measure long before it is landscaping, and letting it bake because well the margin was not there.

Abundance is the word I have been most nervous about using, because it sounds like an invitation to waste, and it is the opposite. It does not mean unlimited consumption and it does not mean careless. It means headroom. It is the ordinary condition of a country that built slightly more capacity than it strictly needed, which is exactly what every generation before ours did with reservoirs, sewers and railways, and which we have somehow come to regard as an indulgence rather than the baseline.

We inherited a system built by people who assumed the country would keep growing and built for the country that would exist after they died. Somewhere along the way we started building only for the country that already exists, and calling that prudence.

And abundance is not a word I invented for this article.

It sits in the same Environment Agency chapter as the figure this piece opened with, a few paragraphs further down, and as far as I can find nobody has reported it. Read the modelling all the way through and there is a second 2055 in it. Under the Agency’s Low water needs scenario, England does not have a shortage at all. It has a surplus of almost 2,320 million litres a day.

Both numbers are in the same document. Same country, same climate models, same people doing the modelling. A deficit of 4,370 at one end and a surplus of 2,320 at the other, with about 6,700 million litres a day between them.

And the good end is not a hidden reservoir. It is the Low water needs case: what you get if demand reduction lands, if the schemes get built, if consumption comes in lower than feared. It is not an alternative to the six jobs. It is what the six jobs produce when they work.

One end of that range is a country rationing between its rivers, its farms and its houses every dry summer, arguing about hosepipes while a full reservoir sits behind an undersized pipe. The other is a country with water to spare. Not one litre of the difference depends on a technology that has yet to be invented.

So the question was never really whether England survives 2055. It is which end of its own regulator’s range we intend to stand at, and who we expect to choose.

Avoiding failure is a plan to survive the century. Headroom is a plan to use it.

What would a solution actually have to do?

We now know there is no silver bullet. We know the six jobs. We know where the existing plan is over-reliant. We know merely closing 4.37bn litres does not create resilience.

So before designing anything, set the rules.

A credible solution has to:

1. Protect the water we already have.
2. Reduce waste without making permanent austerity the strategy.
3. Store enough winter water to survive dry summers.
4. Connect the system so water can actually reach where it is needed.
5. Reuse water that continues to exist during drought.
6. Create genuinely rainfall-independent supply where it makes sense.
7. Build headroom so failure of one project doesn’t break the whole plan.
8. Have somebody accountable for the total, not merely their individual piece.

That is the brief. Part Three is where I try to build it.

One more question

We have tested the obvious answers. Six jobs, no single winner, and a plan that leans on hope more than headroom.

And I want to be straight about where that leaves me, because I started this kinda wanting a villain and did not find one. What I found instead was a country that has almost everything it needs and no mechanism for adding it up. The leaks are known. The reservoirs are consented or nearly. The recycling equipment is ready to go. The law that stops a household being sold anything but drinking water is a definition somebody could rewrite in an afternoon. Every one of those is somebody’s job, and none of them is anybody’s problem.

That is a much more frustrating answer than corruption would have been, and a much more hopeful one. Corruption needs punishing. This needs somebody to own it.

So let me end where most people start.

Most people arrive at this subject through one fact, and it is the right one. Water companies lose about 2.9 billion litres a day, roughly a fifth of everything they put into the pipes, while being asked to explain why you should time your shower.

That anger is correct. I want to be clear about that, because this investigation has spent seven chapters complicating it and I do not want to leave you thinking it was misplaced.

But here is what the leaks turned out to be, once I had followed them all the way down.

The sector promised to cut leakage sixteen per cent by 2025. It delivered nine. Four of them came out of those five years worse than they went in. Nobody hid this. It is published, by the regulator, in a table anyone can download, and it has been sitting there since October.

The same is true of everything else in this article. The reservoirs that will not hold water this decade are on a public list with dates against them. The recycling schemes are in gate submissions. The trunk mains that cannot deliver on a hot evening are known to every engineer who works on them. The white paper that never once says the word potable is a free PDF.

None of this is a secret. That was the thing I kept walking into. There is no cover-up here, and in a way that is worse, because a cover-up you can expose. This is all just sitting in the open, in fifteen different places, and nobody whose job it is has been asked to add it up.

So the leaks are not the story. They are the part of the story you can see from the street.

England is not short of water, or short of engineering, or even short of money when it doesn't want to be. It is short of anybody whose actual job it is to stand back, look at the whole thing, and say: this does not come to enough, and here is what changes.

Enough.