Nothing Ends Gradually

Where the crisis lives

The energy crisis does not live where it is discussed. It is not in the summit communiqué or the photograph on the steps. It lives at the end of a chain, and you are the end of the chain.

Follow anything you bought this week backwards. The bread was baked in an oven, milled by a motor, hauled by a truck and grown with fertilizer made largely from natural gas. The rent was set by a building poured from cement, and cement is limestone cooked in a kiln. Every link in that chain bought energy from the link before it, added its margin, and passed the total forward.

At the end of the chain there is no one left to pass it to. When energy is expensive, life is expensive. Not as a mood. As arithmetic, compounded at every link and settled at the last one.

The teaspoon

Now look at what the world celebrates, and do the arithmetic that almost nobody does out loud.

A ten kilowatt array on a school roof gets a ribbon. A one megawatt solar field gets a press release. A one gigawatt plant gets a national announcement and years of construction. All of it is honest work by serious people. Almost none of it is ever measured against the thing it is meant to fix.

So measure it. In 2025 the world generated 31,747 terawatt hours of electricity External¹. Spread across the year, that is about 1,007 megawatt hours every second Calculated. Now put one plain question to each celebrated project: how long do the world’s grids take to use up one full year of its output?

The rooftop array: about thirteen thousandths of a second Calculated. The solar field: under two seconds Calculated. The gigawatt plant, running at nine tenths of its capacity all year: about two hours and ten minutes Calculated.

Years of work, one afternoon of the world’s demand. That is not a failure of effort. It is a failure of scale, and effort at the wrong scale never adds up to the right one.

We have been bailing an ocean liner with a teaspoon, and holding ceremonies for the teaspoon.

Every nation except two

Here is the number this article exists to deliver, and it is not rounded up for effect.

One CTMP module is designed as 300 gigawatts of firm hydroelectric capacity Target, running above 93 percent of all hours Target. Its modeled net generation is 2,496.6 terawatt hours a year Modeled². Call it 2,500.

Now set it beside the countries of the Earth, using what each one actually generated in 2025 from every source it has: every coal plant, every dam, every reactor, every solar park External¹.

• India, the most populous nation on the planet, generated 2,082 terawatt hours. One module is about 20 percent more Calculated.
• Japan, Germany and Brazil together generated 2,281. One module is more than all three combined Calculated.
• The whole African continent generated about 986 External³. One module is more than two and a half times that Calculated.
• Pakistan generated 177. One module is about fourteen times as much Calculated.

Go down the list of every nation on Earth and the pattern holds almost to the top. One module out-generates every one of them except two: China and the United States Calculated.

Measured against the whole species, one module is 7.9 percent of all the electricity humanity generated in 2025 Calculated. The world’s grids would take about 29 days to use one year of it Calculated.

The gigawatt plant lasts the world one afternoon. The module lasts it a month.

It is designed to deliver that through drought, through winter and through windless weeks, because it does not wait on weather. The largest power station ever built, the Three Gorges Dam on the Yangtze, has 22.5 gigawatts External⁴. A module has more than thirteen times that capacity Calculated. Three Gorges’ best year ever, 2020, produced 111.8 terawatt hours, and its average year is about 95 External⁴. A module is designed to deliver more than twenty-two times that record year Calculated, because a river rises and falls with the seasons and a module is designed not to.

Three Gorges took seventeen and a half years, from the start of construction on 14 December 1994 to its last main turbine in July 2012 External⁴. Hold that number. It is about to matter.

One more figure belongs here, the one that sounds like a misprint. Inside the platform, power moves between its own divisions at a modeled owner cost of 0.08 cents per kilowatt hour Modeled. Everyone else buys it at a posted 2.5 cents Target, a price fixed in the platform’s Stewardship Charter, with a ceiling of 5 cents on any verified end user bill Target⁸.

Why nothing this size gets built

If a machine like that is possible, the obvious question is why no one has built one. The answer is not physics. It is shopping.

Bent Flyvbjerg of Oxford has gathered the outcomes of more than 16,000 large projects across 136 countries External⁵. Of all of them, 8.5 percent finished on budget and on time External. One half of one percent, one project in two hundred, finished on budget, on time and delivered the benefits it promised External. Large dams fare worse. Across 245 of them in 65 countries, actual costs ran an average of 96 percent over estimate, nearly double, and schedules ran 44 percent long External⁶.

Everyone knows big projects fail. Almost everyone draws the wrong conclusion. Size is not what kills them. Method is.

The conventional builder owns nothing and buys everything. The boring machines come from one manufacturer’s order book, the turbines from another’s, the steel from a third, and the money from lenders whose interest runs every day the machines are late. Every seller adds a margin and a queue. The margins compound. The queues compound. The finished thing arrives late, over budget, and belongs to its creditors before it has served a single person.

At the CTMP’s scale that method does not merely slow down. It fails. The platform’s technical paper puts one module’s requirement at about 300 tunnel boring machines and 200 large turbine sets a year Modeled⁷: fleets, not units, set against the annual output of the entire world industry. No one can join the back of a queue for more than the factory behind it can make. For a builder that shops, a project this size is not expensive. It is impossible, and impossibility is not a price anyone can negotiate down.

So the CTMP does not shop. It is designed to make its own tunnel boring machines, turbines, steel, concrete, cable, fuels and chips, on its own power, and to move them between its own divisions at cost Target⁷. The machine factories are the first buildings energized, so machines are built while the first civil works begin. The rock that comes out of the tunnels goes into the concrete and the steel. The workforce lives a train ride from the tunnel face. That is how the published build window is five to six years from first excavation to commissioning Target⁷, set against seventeen and a half years for Three Gorges at less than a thirteenth of the size Calculated.

It is the oldest rule in manufacturing, applied without exception: make the thing yourself, and nobody can charge you for waiting.

The empty seat

Now the part most people read as a footnote. Slow down for it.

Every large machine on Earth answers to whoever holds its paper. A dam built on bonds answers to its bondholders. A listed utility answers to its shareholders. A project built on loans answers to its lenders, who hold the one lever that matters: they can stop the money.

The CTMP was designed so that no one can hold that lever.

It carries zero debt Target⁸: no loans, no bonds, no sovereign guarantees. It will never list on any stock exchange Target⁹. Outside capital is accepted, and the first two or three modules are planned to be built with it Target⁹, but only on terms that can never become control: no single investor may hold more than 35 percent of any vertical Target⁹. After those first modules the platform is designed to finance itself, reinvesting 80 to 90 percent of what it earns Target⁹.

Read that as a question rather than as finance. Who, then, can steer this machine? Not a bank: there is no loan to call. Not a market: there are no listed shares to sell. Not an investor, however large: control is not for sale at any price.

The seat where control normally sits is empty. It was built empty, on purpose.

An empty seat forces the question every serious reader should ask next. If money does not decide where this goes first, what does?

A count.

Not a vote, and not a contract. The Stewardship Charter is precise about what the count can do, and the precision is the point. It describes The Wall as a way to make demand visible, not a way to govern: it binds no one and enforces nothing ⁸. What it does, under the programme’s published rule, is set the order in which countries are formally examined Target⁹, and put public demand in front of the entire world in numbers no one can pretend not to see.

Keep the empty seat in mind. It comes back.

What one module carries

Cheap, firm power is the master input of everything else, and the platform is designed to convert it on purpose. Four conversions matter most.

Water

Each module is designed to give its host nation two billion cubic meters of fresh water a year, with essential use priced at zero Target⁸. Two billion cubic meters is about 5.5 billion liters a day Calculated: 55 liters a day for each of 100 million people Calculated, inside the 50 to 100 liters the World Health Organization gives for basic household needs External¹⁰.

The obvious objection is cost, so do the cost. Modern seawater desalination uses roughly three to four kilowatt hours per cubic meter External¹¹. For two billion cubic meters that is six to eight terawatt hours a year Calculated, between a quarter and a third of one percent of one module’s output Calculated. At the platform’s owner cost that energy is worth $4.8 to $6.4 million a year Calculated. Even at the full public tariff it is worth $150 to $200 million Calculated.

Now the other side of the ledger. Seawater carries about 35 kilograms of dissolved minerals in every cubic meter External¹². Making two billion cubic meters of fresh water leaves at least 70 million tonnes of them behind every year Calculated. Nearly everyone else pumps that brine back into the sea as waste. The CTMP refines it. Its desalination and brine documentation puts the gross value of the refined minerals at an indicated $30 billion per module year Modeled¹³, bounded by how much the world’s markets can absorb.

Put the two sides next to each other. The energy to make the water, even at the public price: under $200 million. What the brine leaves behind: an indicated $30 billion, at least 150 times as much Calculated.

The water is not given away despite the economics. It is given away because of them.

Work

The CTMP’s consolidated workforce model counts two populations and never adds them together ¹⁴. The first is construction: the initial build of one module requires 25.9 million construction job years Modeled. The second is the standing workforce: about 3.0 million positions in Year 1 Modeled, of which 1.04 million are industrial and technical and 1.98 million are the civic and service economy of the first city the platform builds. By Year 30 the standing workforce is about 47.1 million Modeled, of which 15.8 million are industrial and 31.3 million civic. These are central cases with published ranges on either side, and they are requirements, not a hiring plan. No single nation meets them alone.

Terms

The public price is a posted 2.5 cents per kilowatt hour Target, with a ceiling of 5 cents on any verified end user bill Target. The enforcement design contains no weapon: no shutoffs, and hospitals, water systems and life safety loads can never be used as leverage against anyone Target⁸.

Carbon

Here this article will be more careful than most climate announcements, because the platform’s own carbon paper is ². In its current edition, one module’s total avoidance runs from about 0.46 to 1.34 billion tonnes of carbon dioxide equivalent a year across three illustrative cases: low, central and fossil heavy Modeled. It also prints a floor case in which the module’s power displaces no fossil generation at all. There only the platform’s firm industrial substitution survives, the ledger still charges the platform for its own upstream emissions, and the total falls to about 0.02 billion tonnes Modeled. The paper names its own open dependency: the share of power exported to the grid awaits a completed energy balance. Every tonne is to be metered at the point of delivery rather than estimated in a certificate, and the CTMP will never sell a carbon credit Target.

For scale, humanity emitted 42.2 billion tonnes of carbon dioxide in 2025 External¹⁵. One module does not fix that alone, and this article will not pretend it does. What it offers is rarer: a number that prints its own zero, so that anyone can check it.

The one thing it cannot make

The platform is designed to make things abundant: power beyond a nation’s use, water beyond its thirst, work beyond its young. There is exactly one thing no engineering can multiply.

Order.

Someone is examined first. Someone’s 50 megawatt proof of concept is poured first. Someone’s first module switches on while everyone else’s is still a drawing. And the head start does not stay the same size.

The design reinvests to grow capacity about ten percent a year Target⁹. Run that on two nations. The first switches on today. The second switches on ten years later. On the day the second nation starts, the first nation’s capacity has already grown to 2.59 times its starting size Calculated. Ten years after that, the first stands at 6.73 times and the second at 2.59 Calculated. For as long as both grow at the same rate, the ratio between them never closes, and the gap between them in gigawatts widens every year Calculated.

A decade is not a delay. It is a permanent ratio.

Every industrial revolution has had this property, and in every one the order was set by capital, by empire or by force. The results are still on the books. In 2023, 666 million people had no electricity at all External¹⁶. In sub-Saharan Africa the number was 566 million in 2010 and 565 million in 2023 External¹⁶: thirteen years of new connections, cancelled almost exactly by the growth of the population still waiting. Africa is home to 18.8 percent of humanity External¹⁷ and generates 3.1 percent of its electricity External³. That is what the order of arrival looks like when it is set by who can pay first.

This time the order is set by a count.

Under the programme’s published rule, every country has its own required count: five percent of its population, never fewer than 250,000 Target⁹. Countries are ranked by progress toward their own number, not by raw totals, so a nation of ten million climbs a mountain of its own size, not India’s. The first country to reach its count earns the first formal review Target. A review is not a deployment. It examines the coastline, the ground, the grid and the host government’s own answer. If the answer is yes, the build begins with a 50 megawatt proof of concept, then 250 megawatts, then module by proven module Target⁹.

The most valuable position of the coming century, first in that order, is currently priced at thirty seconds of an ordinary person’s attention. It is priced that low for one reason only: most of the world has not looked yet.

The hundred in the square

If you have read this far, you are not a skimmer, so here is the layer underneath everything above.

The energy crisis was never really about energy. Energy is the carrier. For five centuries, in system after system, a thin layer has stood between ordinary people and the essentials of life, charging for the crossing. It has rarely needed force to stay there. It has needed one thing only: that the people on the paying side could not see each other. Each wanted the change. Each assumed they were nearly alone. So each stayed where they were.

In 1978 the sociologist Mark Granovetter wrote down the arithmetic of that silence External¹⁸. Picture a hundred people in a square. Each carries a private number: how many others they must see act before they will act themselves. One person’s number is zero, so they will act alone. One person’s number is one. One person’s is two. And so on, up to ninety-nine.

The person whose number is zero acts. That makes one, which moves the person whose number is one. That makes two, which moves the person whose number is two. The chain runs, one link at a time, until all hundred have moved.

Now change a single person. Replace the one whose number was one with someone whose number is two. Everything else stays identical. The crowd’s average number moves from 49.50 to 49.51 Calculated.

The person whose number is zero acts. Then nothing happens. The next person in line sees one and needs two. No one’s number is one. The chain breaks at its second link, and ninety-nine people, every one of whom moved in the first crowd, stand still and watch one person act alone.

Same crowd. Same wants. Opposite outcomes. Granovetter’s warning was that groups with nearly identical preferences can produce opposite results, so no one can read what people want from what a crowd does.

The silence on the paying side was never consent. It was a chain with a missing link.

That is what The Wall is: a coordination instrument disguised as a counter. It makes the wanting visible without making the wanter visible. No name, no email, no account, no money. One human, one voice: a head of state and a night shift nurse count exactly the same. And every voice it records becomes the number someone else was waiting to see.

On 3 August 2026 The Wall held 50,064 voices Measured¹⁹. On 29 September 2026 it held 667,101, from 176 countries Measured¹⁹. The fortress was never built of stone. It was built of gaps, and the count is how the gaps close: one person at a time, in public.

What this article does not claim