The Engineered End of Scarcity

1. Introduction

1.1 The claim

The CTMP rests on one claim. When firm baseload electricity is produced at a modeled internal transfer price of 0.08 cents per kWh MOD, sold at a posted tariff of 2.5 cents per kWh TGT, converted inside a vertically integrated industrial platform, and governed so that it cannot be captured, the physical cost floor of modern life moves. The claim has three parts, and each is necessary. Cheap power without conversion has nowhere to go: outside China and the United States, no national electricity system is as large as the output of one module CALC (Section 6). Conversion without governance becomes a new concentration of rent: the same machinery that lowers the cost of water, steel and computation can be used to raise a toll on them. Governance without a public mandate becomes a negotiation between capital and ministries in rooms the public does not enter. The platform is therefore designed as one system in which each part supplies a condition that the others require. This paper sets out that system as a chain of propositions, each of which can be checked against a cited source, a stated equation or a published rule.

1.2 Why the energy layer

Every physical good carries an energy cost, and most of that cost is invisible to the person who pays it. Electricity appears on a bill once. It appears again inside the price of water, where desalination and pumping are energy processes; inside fertilizer, where natural gas is the principal input to ammonia; inside steel and cement, two of the most energy-intensive materials in use; inside freight, refrigeration, construction and rent; and, increasingly, inside computation, where data centers consumed an estimated 415 TWh in 2024 EXT [3]. When the price of energy rises, these costs rise together. When the supply of energy is unreliable, or cannot be connected, the goods that depend on it are rationed by queue rather than by price. Section 3 quantifies these channels. The result established there is narrow and it carries the rest of the argument: the energy layer is the common term in the cost of physical life, and a change in that term propagates into every good that contains it.

1.3 The argument in seven links

The argument of the paper is a chain. Each link is established in a named section and depends on the link before it.

1. The energy layer sets the cost floor. The price, firmness, financing and governance of energy determine the cost and availability of water, food, materials and computation (Section 3).
2. One module changes the scale of the layer. A 300 GW module of firm baseload power TGT produces at least 2,444 TWh per year CALC, a quantity comparable to national electricity systems (Section 4).
3. At the Charter's prices, energy stops being the binding cost. The posted tariff of 2.5 cents per kWh TGT is 31 percent of the 2024 US industrial average price CALC, the internal transfer price is about one percent of it CALC, and, on costs specified independently of the CTMP model, the tariff holds for capital overruns of several times the planning figure (Section 5).
4. Output at that scale must be converted where it is made. One module exceeds the total generation of every country except two CALC. No host grid can absorb it, so the platform must carry its own load (Section 6).
5. Seventeen integrated verticals are that load. They turn power into water, steel, concrete, fuels, chemicals, computation and the machines that build further modules (Sections 7 and 8).
6. Reinvestment turns the platform into a compounding system. Each module's revenue funds further capacity, and each year's materials build the next year's plant (Sections 9 and 10).
7. Because the system is powerful, allocation and governance are part of the machine. The Wall decides where review happens, on a public record with fixed clocks; the Stewardship Charter decides what the platform can never become (Sections 11 and 12).

The consequences for health and climate follow from the chain rather than from separate claims (Section 13). The gates that each site must pass are listed in Section 14. Section 15 discusses what the arithmetic shows, Section 16 states the concessions, and Section 17 concludes.

1.4 What this edition changes

This edition replaces the paper of 2 May 2026 [1]. It updates every external statistic to the latest published year, mostly 2025. It corrects the share of world electricity represented by one module from "roughly one-tenth" to 7.7 percent CALC. It adopts the review procedure published in Order of Operations [2], which separates the first count that earns First Independent Review from the full count that opens the formal deployment process, and which replaces the earlier status vocabulary with two public clocks and seven statuses. It adopts the current growth target of 25 percent per year TGT and the reinvestment shares of 50 to 60 percent of revenue for the first module and 80 to 90 percent for later modules TGT. It adds four analyses that the earlier edition did not contain: the absorption problem, a sensitivity analysis of the cost floor, a mass balance of brine minerals against world markets, and yield arithmetic for the principal verticals. It removes mortality figures pending publication of their methodology, and it states the platform's position on carbon credits. Appendix B lists every change, including the corrections.

2. Methods and conventions

2.1 Classification labels

The five labels defined at the head of this paper are applied to every quantitative claim. Three conventions govern their use. First, numbers fixed by the Stewardship Charter or by the published review procedure, such as the posted tariff, the Verified resale ceiling, the first-count and full-count formulas and the two clocks, carry the Target label in the sense of a binding commitment; they are rules, not estimates. Second, a calculated value inherits the status of its inputs. When this paper states that one module equals 7.7 percent of world electricity demand CALC, the numerator is a design value and the sentence is a statement about the design. Third, where a calculation uses a parameter chosen by this paper rather than taken from CTMP or from a cited source, for example a reference volume of 100 liters per person per day, the parameter is marked as such in Table 2, and every result that uses it is labeled Calculated.

2.2 Sources and vintage

External statistics are taken from the most recent edition of each source available on 16 October 2026. Electricity statistics are for calendar year 2025 from Ember [4, 5]; energy-related emissions for 2025 from the International Energy Agency [6]; data center electricity from the IEA [3, 7] and Lawrence Berkeley National Laboratory [8]; access to drinking water and sanitation for 2024 from the WHO and UNICEF Joint Monitoring Programme [9]; access to electricity for 2024 from the Tracking SDG 7 report [10]; steel for 2025 from the World Steel Association [11, 12]; hydrogen for 2025 from the IEA [13]; minerals for 2025 from the US Geological Survey [14]; generation costs from IRENA [15] and Lazard [16]; and US delivered prices for 2024 from the Energy Information Administration [17]. Datasets from the same publisher can differ slightly because of revisions; where two figures exist for one quantity, this paper uses the figure in the publisher's headline release and names it. CTMP parameters are taken from the published CTMP record [18, 2, 19] and from the stated parameters of the platform's financing plan. peoplesctmp.org is the canonical source for all of them.

2.3 Equations

All derived values in the paper follow from twelve relations. Symbols are defined where they first appear and collected in Appendix C.

Annual energy E from net capacity P, the 8,760 hours of a year h, and capacity factor CF.

The standard hydroelectric relation: power from machine efficiency η, water density ρ, gravitational acceleration g, volumetric flow Q and net head H. Stated to identify the governing quantities; no values are disclosed (Section 2.6).

Unit cost c of power from capital K, the capital recovery factor at rate r over n years, and annual operating cost as a fraction f_OM of capital. Where r = 0 the platform's annual charge is taken directly from the internal transfer price (Section 5.2).

Total interest I paid on capital K financed entirely by debt with level annual repayment.

Capital reinvested per year C_k by module k, from its reinvestment share of revenue φ_k, the share of output sold externally s, the posted tariff T and annual output E; and the same quantity in Core-equivalents M_k at the Core planning capital K.

Years for an absorbed share x of output to grow to full output at a constant annual growth rate γ.

Output after t years of constant growth γ, and the doubling time t₂.

Energy E_w to desalinate volume V at specific energy e_w, and the energy cost of one cubic meter c_w at electricity price p.

Seawater feed V_f for product volume V at recovery ratio R, and the mass m_i of constituent i contained in the feed, from seawater density ρ_sw and mass fraction w_i.

Annual output Y_j of product j from a share f_j of annual energy at specific energy e_j.

Gross annual displacement D when a share f_d of output replaces generation of lifecycle intensity ε_d.

The first count N₁ and the full count N₂ of The Wall for a country of population N_pop, and its completion κ from v recorded voices [2].

2.4 Parameters

2.5 Units and rounding

Energy is given in terawatt-hours (TWh; one TWh is one billion kWh); power in gigawatts (GW); mass in megatonnes (Mt) and gigatonnes (Gt); water in cubic meters. Prices are in US cents per kWh or US dollars, nominal. Every value is computed at full precision and rounded for presentation, usually to three or four significant figures; ratios are given to two or three significant figures. Rounded values may therefore not add exactly. The complete set of derived values, with inputs, is in Appendix A.

2.6 Disclosure boundary and exclusions

Two exclusions apply. First, the paper withholds the hydraulic particulars of the Core (flow, head, geometry, machine counts and setpoints) under the disclosure note. Equation 2 is stated so that the reader can see which physical quantities govern output, not to imply values. Second, the paper reports no mortality figures. CTMP's mortality estimate, which sets the pace of the project, is withheld until its full methodology can be published and checked [18]. Until then, health is described through exposure statistics (Section 13). Those are counted by the cited agencies and do not depend on attribution models.

3. Scarcity is engineered through the energy layer

Link 1 of the argument: the price, firmness, financing and governance of energy determine the cost and availability of the goods that contain it. This section quantifies five channels through which that happens and closes with the scale of the change now under way.

3.1 Energy as the master input

Water. The minimum theoretical energy required to separate fresh water from seawater is about 1.06 kWh per cubic meter at 50 percent recovery EXT [26]. The most efficient large plants now operate at 2.34 kWh per cubic meter overall EXT [20], and 3 to 4 kWh per cubic meter is typical of modern seawater reverse osmosis EXT [21]. Because separation itself sets a floor on energy, desalinated water is physically a product of electricity, and its price follows the price of power more closely than it follows any other input. In the most recent global inventory, the world's desalination plants produced about 95.4 million cubic meters of fresh water per day and discharged about 141.5 million cubic meters of brine EXT [21].

Food. Ammonia is the starting point of all mineral nitrogen fertilizer, and about 70 percent of the world's ammonia is used to make it EXT [31]. Just over 70 percent of ammonia is produced from natural gas, consuming about 170 billion cubic meters of gas per year, and production emits about 450 Mt of CO2 directly, around 2.4 tonnes per tonne of ammonia EXT [31]. The cost of feedstock gas is more than half of the levelized cost of conventional ammonia EXT [27]. A share of the price of bread is therefore a gas price, transmitted through fertilizer.

Steel. The world produced 1,849.4 Mt of crude steel in 2025 EXT [11]. The industry's reported average energy intensity is 20.95 GJ per tonne and its CO2 intensity 1.92 tonnes per tonne EXT [12]. Steel is the skeleton of every power station, port, transmission line, vehicle and building, so its price enters every other physical system.

Cement and concrete. About 60 percent of the cement industry's CO2 comes from calcination, the chemical release of carbon dioxide from limestone, and the remainder from fuel and electricity EXT [29, 28]. The distinction matters in Section 7.4: cheap power removes the energy share of cement's burden but not the process share, which requires a change in binder chemistry.

Computation. Data centers used about 415 TWh of electricity in 2024 EXT and an estimated 485 TWh in 2025 EXT. The IEA base case projects about 945 TWh in 2030 EXT [3, 7]. In the United States, data centers used 176 TWh in 2023, 4.4 percent of national electricity, and are projected to reach 6.7 to 12 percent by 2028 EXT [8]. The newest large consumer of electricity is arriving on grids that were already constrained.

3.2 Transmission into prices

When energy prices move, other prices move with them. In October 2022 the annual inflation rate of the euro area reached 10.6 percent EXT. Energy alone contributed 4.44 percentage points EXT, 41.9 percent of the total CALC, and energy prices stood 41.5 percent higher than a year earlier EXT [30]. The contribution recorded under "energy" is only the direct part. Food, services and industrial goods, which contributed most of the remaining points, contain energy in their production and transport, and that content is not separated in the index. The direct share is therefore a lower bound on the role of energy in that episode. The mechanism generalizes: a household pays for energy several times, once on its own bills and again inside everything it buys.

3.3 The finance layer

The cost of electricity from a new plant is largely a cost of capital, and capital is not priced equally across the world. In emerging and developing economies the cost of capital for a typical utility-scale solar plant has been two to three times that in advanced economies and China, and financing accounted for around half of the levelized cost of such a plant, against 25 to 30 percent in advanced economies EXT [32]. Emerging and developing economies outside China hold two thirds of the world's population and receive less than one fifth of its clean energy investment EXT [32]. The external debt of low and middle income countries reached a record 8.9 trillion US dollars in 2024, their interest payments reached a record 415.4 billion US dollars, and between 2022 and 2024 they paid 741 billion US dollars more in debt service than they received in new financing EXT [33]. In such economies the price of a kilowatt-hour is, to a large extent, the price of money. A platform that carries no debt removes that term from its cost. Section 5.4 measures the size of the term for one module.

3.4 The connection queue

Scarcity can take the form of a queue rather than a price. At the end of 2024 about 2,290 GW of generation and storage capacity, comprising 1,400 GW of generation and about 890 GW of storage, was waiting to connect to the US transmission grid EXT [34]. Of the capacity that requested interconnection between 2000 and 2019, only 13 percent had reached commercial operation by the end of 2024, and 77 percent had been withdrawn EXT [34]. A factory, a data center or a desalination plant that cannot connect is rationed whatever it is willing to pay. For loads on its own site, a platform that generates its own firm power is not in that queue; its exports to host and neighboring grids still require interconnection.

3.5 The access deficit

At the end of the chain are people without the services that energy makes. In 2024, 655 million people had no access to electricity, 563 million of them in Sub-Saharan Africa EXT [10]. In the same year 2.1 billion people lacked safely managed drinking water, including 106 million who drank untreated surface water, and 3.4 billion lacked safely managed sanitation EXT [9]. In 2022 the WHO reported that 99 percent of the world's population breathed air exceeding its air quality guideline limits EXT [35]. These are counts of exposure, not estimates of harm, and they are reported here as such (Section 2.6).

3.6 The scale of present change

The current transition is real and it is measured. World electricity demand rose by 849 TWh in 2025, or 2.8 percent, to a record 31,779 TWh EXT [4]. Solar generation grew by 636 TWh, the largest single-year increase of any source on record apart from the post-pandemic rebound of coal in 2021; wind and solar together grew by 841 TWh; renewables reached 33.8 percent of generation and overtook coal, at 33.0 percent, for the first time in a century; and clean sources met all of the growth in demand, so that fossil generation fell by 38 TWh, or 0.2 percent EXT [4]. Fossil fuels still supplied about 57 percent of world electricity, 18,219 TWh EXT [4, 5], and world energy-related CO2 emissions reached a record of nearly 38.4 Gt EXT [6].

The arithmetic of scale is plain. Solar's record growth in 2025, the largest increase of any source that year, equals 26.0 percent of the design output of one CTMP module CALC; the combined growth of wind and solar equals 34.4 percent CALC. The transition is adding capacity at the margin of a system whose base is still fossil. The CTMP is not an alternative to that transition and does not compete with it. It operates at a different scale and on a different variable: firm output, delivered at a fixed price, at the site where it is converted.

3.7 Synthesis

The five channels share one structure. In each of them the binding constraint on a physical good is not the availability of matter: seawater, iron ore, limestone, atmospheric nitrogen and silicon are abundant. The constraint is the price of energy, the cost of the capital needed to produce it, the queue for connecting it, or the arrangements that decide who receives it. All four are produced by institutions, contracts and financing structures. They are engineered in the ordinary sense of the word: designed, built and maintained. What is engineered can be engineered differently. The CTMP acts on the layer where the constraint is produced. It supplies firm energy at a fixed price, without debt, at the site of conversion, under rules that prevent whoever controls the energy from pricing it upward. The remaining sections test whether the arithmetic supports that design.

4. The Core: firm baseload at module scale

Link 2: one module changes the scale of the energy layer. This section states the design basis, places it against national systems and against the largest hydroelectric project under construction, and sets out the published proof ladder.

4.1 Design basis

The CTMP Core is a subsurface, ocean-fed hydroelectric architecture. It uses the ocean as its source reservoir, conveys water through engineered rock corridors and buried or submerged hydraulic structures to a distributed machine hall, converts it to electricity, and returns it to the ocean. It has no surface reservoir, no visible dam, no lagoon and no unnecessary above-ground exposure. The first canonical module is designed for 300 GW of net firm baseload TGT at a capacity factor above 93 percent TGT [18]. Equation 1 gives its annual output.

At a capacity factor of 0.95 the output is 2,497 TWh CALC. This paper uses 2,444 TWh throughout, the value at the lower bound of the design range, so that every ratio derived from it understates the design rather than overstating it. "Firm" means that the output is available continuously, independent of weather and season, and can be scheduled. It is the property that makes the module a base for industry rather than a supplement to a grid.

4.2 The governing relation

The output of any hydroelectric machine is governed by Equation 2: power is the product of machine efficiency, water density, gravitational acceleration, volumetric flow and net head. The relation has been the basis of hydroelectric engineering for more than a century, and it is the relation that governs the Core. What is particular to the Core is the disciplined application of established hydraulic, geotechnical, mechanical, electrical and control engineering at a scale matched to the problem, with the ocean as the reservoir. The specific values of flow, head and geometry are withheld under the disclosure note. Their demonstration belongs to the proof ladder (Section 4.6) and to independent engineering review (Section 14).

4.3 Scale in context

Figure 2 places one module against the total electricity generation of the largest national and regional systems in 2025 EXT [5]. One module generates 1.17 times the total of India CALC, 2.05 times that of Russia CALC, 2.37 times that of Japan CALC and 3.25 times that of Brazil CALC. It generates 1.55 times the whole Middle East CALC, 1.78 times ASEAN CALC, 1.32 times Latin America CALC and 2.45 times the whole of Africa CALC. It equals 88 percent of the generation of the European Union CALC, 54 percent of that of the United States CALC and 23 percent of that of China CALC. Only two countries generate more electricity than one module CALC. Against world aggregates, one module equals 7.7 percent of world electricity demand in 2025 CALC and 55 percent CALC of all hydroelectric generation on Earth, which was 4,433 TWh in 2025 EXT [5].

Figure 3 compares the module with annual flows rather than totals. One module's output is 2.88 times the entire growth in world electricity demand in 2025 CALC, 5.04 times the electricity used by all of the world's data centers in 2025 CALC, and 2.59 times the IEA base-case projection for data centers in 2030 CALC. These comparisons do not imply that the module serves any of those uses. They establish an order of magnitude: one module is a national-scale energy system delivered at one site.

4.4 Firmness

Capacity factor measures how much of the year a plant delivers its rated output. The US nuclear fleet, among the most consistently operated large fleets in the world, averaged 91 percent in 2025 EXT [37]. The design capacity factor of the Core, above 93 percent TGT, places it in that class. Variable sources can be cheap at the point of generation and still impose costs elsewhere in a system: storage, curtailment, backup capacity and transmission. Lazard's 2025 analysis puts the cost of variable output once firmed, including the cost of the backup capacity it requires, at between 71 US dollars per MWh, for unsubsidized wind in the US Midwest, and 164 US dollars per MWh, for solar with storage in the mid-Atlantic region EXT [16]. A firm source avoids most of that charge, but not all of it. Planned maintenance and forced outages must still be covered by reserve capacity, and concentrating 300 GW at one site makes the size of the largest single loss a design question; the distributed machine hall (Section 4.1) and the redundancy described below are intended to keep any single loss a small fraction of output. This does not make solar, wind, storage or nuclear power competitors of the Core. It places the Core in a different category: a continuous industrial platform that can support other systems rather than depend on them for its own availability.

Reliability in the Core is a design property, engineered through redundancy, isolation, protection, sectionalization, black-start capability and deterministic enforcement of operating rules. The public statement of the principle is sufficient here. Failure should isolate locally rather than propagate. Critical loads keep priority continuity. Failure of an external grid should not collapse the platform. The platform should be able to help restart a grid rather than wait for a grid to restart it. The detailed protection architecture remains internal.

4.5 The contemporary comparator

The largest hydroelectric project under construction anywhere is the Medog cascade on the lower Yarlung Tsangpo in Tibet: 60 GW EXT, a planned output of 300 TWh per year EXT, and a reported investment of 1.2 trillion yuan, between 167 and 176 billion US dollars at the exchange rates used in reporting EXT [36]. Construction began on 19 July 2025. Its implied capacity factor is 57.1 percent CALC, typical of a river system whose output follows the hydrological year. Against Medog, one CTMP module has 5 times the capacity CALC and 8.15 times the annual energy CALC for 0.71 to 0.75 times the planning capital CALC. Capital per unit of capacity is 417 US dollars per kW for the module CALC, against 2,783 to 2,933 US dollars per kW for Medog CALC. Capital per kWh of annual output is 10.9 to 11.5 times lower CALC.

Of that last ratio, a factor of 1.63 comes from capacity factor alone CALC: an ocean reservoir has no dry season. The remainder is what the platform's design asserts: a supply chain that makes its own machines, steel, concrete and power at cost (Section 7.7); no debt, and therefore no interest during construction (Section 5.4); and continuous construction at one site. The comparison is given so that a reader can see exactly what the planning figure states, and so that the 50 MW and 250 MW stages can be read against it.

4.6 The proof ladder and the build time

Physical claims meet physical tests in stages: a 50 MW proof of concept TGT, then 250 MW TGT, then module by proven module [2]. The 50 MW proof of concept is complete in design, and its first physical construction is the next milestone [18]. Figure 4 shows the ladder at true scale. The first step is a factor of 5 CALC; the second is a factor of 1,200 CALC. The published build time of a module, from first excavation to switch-on, is 5 to 6 years TGT [18]. That schedule depends on building the machines on site while the ground is prepared, rather than waiting in the order books of external suppliers; Section 7.7 describes the mechanism. The platform has bound itself to metered, public reporting from the first delivered kilowatt-hour [18]. Every Core figure in this paper is a design value until that reporting replaces it. The labels say so, and the ladder is how the replacement happens.

5. The cost floor

Link 3: at the Charter's prices, energy stops being the binding cost of the goods that contain it. This section defines the three prices of power in the platform, derives the annual charge implied by the internal transfer price, tests the posted tariff against capital overruns and debt financing, and shows how reinvestment turns the margin into new capacity.

5.1 Three prices

Power in the platform carries three prices, each with a distinct function. The internal transfer price, 0.08 cents per kWh MOD, is the computed cost at which power moves from the Core to the other verticals. It is an internal model value based on zero-debt financing, a high capacity factor, vertical integration and a very large delivered-energy denominator; it is not a regulated benchmark or a public-utility rate filing. The posted tariff, 2.5 cents per kWh TGT, is the price at which the Core sells power at the metered delivery point, to utilities and other buyers. It is fixed by the Stewardship Charter and no operator can raise it [18]. The Verified resale ceiling, 5 cents per kWh all-in TGT, caps the price at which power may be resold under the CTMP Verified label [18]. The posted tariff is 31.25 times the internal transfer price CALC. Section 5.6 sets out what the difference pays for and why it is not a license for extraction.

5.2 The annual charge implied by the internal price

Equation 3 relates the unit cost of power to capital, the rate at which capital is recovered, operating cost and annual output. Read in reverse, it shows what the internal transfer price implies. At 0.08 cents per kWh MOD and 2,444 TWh per year CALC, the total annual cost of the Core is 1.96 billion US dollars CALC, which is 1.56 percent of the 125 billion US dollar planning capital TGT per year CALC. This is the annual allowance the CTMP model makes for capital recovery, maintenance and replacement together. It is low because two terms that dominate the cost of conventional projects are absent. There is no interest, because there is no debt. There is no return requirement on the capital of the Core, because the Core is not financed by instruments that demand one. Long-lived civil works, a continuous capacity factor and the size of the denominator account for the rest. Whether 1.56 percent of capital per year covers the plant's actual recovery, maintenance and replacement over its life is a property of the CTMP model, whose internals are not reproduced here. The derivation shows what the model implies; it does not show that the allowance is sufficient. Section 5.3 therefore tests the posted tariff with costs specified independently of the model. At the posted tariff the modeled annual cost of the Core is 3.2 percent of the revenue the Core would earn if its entire output were sold externally CALC.

5.3 Sensitivity to capital overrun and financing

A planning figure is not a contract price, and the base rates for large projects are severe. In the largest database of major projects yet assembled, covering more than 16,000 projects, 8.5 percent were delivered on budget and on time, and 0.5 percent on budget, on time and with the benefits promised EXT [38]. For large dams specifically, a study of 245 projects built between 1934 and 2007 found actual costs on average 96 percent above estimates, with a median of 27 percent, and three in four over budget EXT [39]. The useful test is therefore not whether the planning figure is exact but how far it can be wrong before the posted tariff fails. Table 3 and Figure 5 answer that question for four financing cases and four levels of capital overrun. Case A holds the annual charge at the modeled 1.56 percent of capital and scales it with capital; it shows what the CTMP model implies and is not an independent test. Cases B, C and D specify capital recovery and operating cost independently of the model, using debt-style recovery that the Charter excludes, to show the result if the Core were financed and maintained as conventional public, commercial or merchant infrastructure. The last column of the table raises the operating, maintenance and replacement allowance in cases B to D to 2.5 percent of capital per year.

Three results follow. First, Case A is conditional: if the modeled allowance of 1.56 percent of capital per year holds as capital rises, the posted tariff is not reached until capital is 31.3 times the planning figure CALC. That is the model's implication, and it is only as strong as the model. Second, the independent cases do not depend on the model. With conventional capital recovery, the tariff is reached at 8.64, 5.33 and 4.04 times the planning capital in cases B, C and D CALC, and at 6.83, 4.58 and 3.73 times when the operating, maintenance and replacement allowance is raised to 2.5 percent of capital per year CALC. The most adverse of these, merchant financing at 10 percent over 30 years with the higher allowance, reaches the tariff at 3.73 times the planning capital, 1.9 times the mean ratio of actual to estimated cost recorded for large dams CALC. Third, of the sixteen combinations in the first five columns of Table 3, fifteen lie below the tariff CALC; the single exception is the merchant case at five times the planning capital. A lower capacity factor changes these results in proportion: at 80 percent instead of 93 percent, the commercial case at the planning capital costs 0.55 cents per kWh CALC.

The conclusion is conditional and bounded. On costs specified independently of the CTMP model, and under financing that the Charter forbids, the unit cost of Core power stays below the posted tariff for capital overruns of up to 3.7 to 8.6 times the planning figure, depending on the case CALC. That range lies above the mean overrun recorded for large dams. It is not a guarantee against every possible overrun, and it covers the Core only: the capital of the verticals is a separate question, taken up in Section 5.7.

5.4 The cost of debt that is not incurred

Equation 4 gives the total interest paid on capital financed entirely by debt with level repayment. Applied to the 125 billion US dollar planning capital TGT, the interest would be 75.0 billion US dollars at 4 percent over 25 years, 119.5 billion at 6 percent over 25 years, 208.1 billion at 8 percent over 30 years and 272.8 billion at 10 percent over 30 years CALC (Table 4). At 6 percent over 25 years the lenders would receive 96 percent of the cost of the build a second time, as interest CALC; at 10 percent over 30 years they would receive 2.18 times it CALC. In the economies where capital costs two to three times as much as in advanced economies EXT (Section 3.3), the higher rows are the ordinary case. Under the Charter none of this is paid. Of the costs quantified in this paper, the interest a debt-financed Core would carry is the largest that the Charter removes, and its removal is a rule, not a forecast.

5.5 Comparison with the market

Figure 6 places the three CTMP prices against generation costs and delivered prices. The posted tariff of 2.5 cents per kWh TGT is 32 percent CALC below the lowest bound of any new-build technology in Lazard's 2025 unsubsidized US ranges, which is 3.7 cents for onshore wind EXT [16]. It is below IRENA's 2024 global weighted averages for new onshore wind at 3.4 cents, solar PV at 4.3 cents and hydropower at 5.7 cents EXT [15]. Against delivered prices, it is 30.8 percent CALC of the average price paid by US industrial customers in 2024, which was 8.13 cents EXT [17], and 19.3 percent CALC of the US average across all sectors, 12.94 cents EXT. The internal transfer price is 0.98 percent of the US industrial average CALC. Generation costs and delivered prices are different quantities, the first excluding transmission, distribution and retail margins, so the figure groups them separately; each comparison holds within its group.

5.6 What the margin is for

The difference between the internal transfer price and the posted tariff is not a license for extraction. It is the operating room that pays for maintenance, resilience, reinvestment, the host country's free water, industrial expansion and long-horizon stewardship. Two rules keep it from becoming a toll. Inside the platform, verticals receive power at the computed internal price, with no margin added between them (Section 12.7). Outside it, the posted tariff cannot be raised and Verified resale is capped at 5 cents per kWh all-in TGT. Utilities that buy at the posted tariff resell within that ceiling, and output that cannot meet the verification conditions cannot be represented as CTMP Verified. The arrangement is truth in labeling for a governed supply chain, not price control. If an intermediary adds an extreme margin, the public record shows that the power is no longer being delivered under CTMP Verified conditions.

5.7 Capital formation and reinvestment

The financing plan has two phases TGT. The first two or three modules TGT are funded by outside capital on the Charter's terms, which admit no debt and no liens. After that, the plan is for the platform to finance itself TGT. Reinvestment is fixed as a share of revenue: 50 to 60 percent for the first module, and 80 to 90 percent for every later module TGT. Equation 5 converts those shares into capital per year and into Core-equivalents, that is, multiples of the 125 billion US dollar planning capital of one Core. At full external sale of output at the posted tariff, power revenue is 61.1 billion US dollars per Core per year CALC. Table 5 gives the result.

Three operating modules, the first reinvesting 50 to 60 percent and two later modules reinvesting 80 to 90 percent, reinvest 128.3 to 146.6 billion US dollars per year at full external sale CALC: the planning capital of about one further Core every year CALC. Three qualifications bound the result. First, every value scales with s, the share of output sold externally at the tariff: output transferred to the verticals at the internal price earns little power revenue but becomes steel, water, fuels and computation, whose external sales at about 20 percent below market TGT are not modeled here. Second, revenue in the first years of a module is limited by the rate at which its output can be absorbed, which is the subject of Section 6. Third, the planning capital covers the Core only. The verticals that absorb a Core's output carry their own construction and operating costs, which are additional and are not modeled here. The identity shows the scale of capital that power revenue can supply once a small number of modules operate. It does not by itself establish that the platform as a whole finances itself: that requires a combined model of the cash flows of the Core and the verticals, including the verticals' construction costs and operating needs. Self-financing after the first two or three modules is the financing plan's Target TGT, and the combined model is its test.

6. The absorption problem

Link 4: output at module scale must be converted where it is made. This section shows why, and sets out the three channels through which the platform does it.

6.1 A module exceeds its host

One module generates more electricity than any national system except those of China and the United States CALC (Figure 2). For almost every possible host country, the output of one module is larger than the country's entire electricity system; it is 2.45 times the generation of the whole of Africa, 1.55 times that of the Middle East and 1.78 times that of ASEAN CALC. A host grid of that size cannot take that power, and a plant whose output cannot be delivered earns nothing and serves no one. Absorption, not generation, is therefore the first practical constraint on the value of a module, and the architecture of the platform follows from it.

6.2 Three channels

Output leaves the Core through three channels (Figure 7). The host grid takes what the host country needs. Surplus power goes to neighboring countries at the posted tariff of 2.5 cents per kWh, over transmission lines that CTMP builds and owns TGT. The remainder is converted on site by the platform's own verticals into water, steel, concrete, fuels, chemicals, computation and machines. Water follows the same rule as power: the host country takes what it needs free of charge, and surplus water goes to neighboring countries, also free of charge TGT. The three channels are not alternatives. They operate together, and their proportions change as host demand, export links and verticals grow.

6.3 The transmission requirement

The Changji to Guquan line in China, which set world records for voltage, capacity and distance when it entered service, carries 12 GW over 3,293 km at ±1,100 kV EXT [25]. Exporting the full output of one module would require 25 lines of that capacity CALC; exporting 100 GW would require 8.3 CALC. Transmission at that scale cannot be bought from the existing market on the platform's schedule. That is why high-voltage direct current is a vertical of its own, and why that vertical also manufactures the platform's transformers TGT.

6.4 The ramp

Absorbed load grows as the host grid connects, export lines are built and verticals come into production. Equation 6 gives the time for an absorbed share x of output to grow to the whole of it at a constant growth rate. At the platform's growth target of 25 percent per year TGT, absorption grows from 10 percent of output to all of it in 10.3 years CALC, from 20 percent in 7.2 years CALC, from 30 percent in 5.4 years CALC and from 50 percent in 3.1 years CALC (Table 6). The 5 to 6 year build time of the Core TGT must therefore also be the build time of its load: verticals, export lines and connections are built in parallel with the Core, so that a substantial share of output can be absorbed at switch-on.

6.5 Consequence

The seventeen verticals are not a portfolio of businesses attached to a power plant. They are its load. Without them, a module is a national-scale generator with no customer large enough to take its output. With them, the output becomes water, materials, fuels, computation and machines, much of which builds the next module. Cheap power alone is not enough, because power that cannot be absorbed has no value. Conversion is what turns a low-cost energy layer into a lower cost floor for physical life.

7. Seventeen verticals as one industrial metabolism

Link 5, first part: the verticals turn power into goods and into the machines that build further capacity. This section gives the registry, the structure of the flows between verticals, and the arithmetic of the principal conversions.

7.1 The registry

The platform has seventeen canonical verticals TGT. Table 7 gives each one's function and its principal internal customers. Numbers in the last column refer to the first column.

7.2 The structure of the flows

The verticals are linked by flows of power, water, materials, equipment and services. The most important of these flows form a closed loop, shown in Figure 8. Core power makes hydrogen; hydrogen reduces iron to steel; steel becomes tunnel boring machines, turbines and HVDC equipment; those machines excavate and build; excavated rock becomes aggregate for concrete; and concrete and machines become further Core capacity. Each step of the loop passes through a good that a conventional project buys from an external market. Each one internalized removes an external price, an external delivery schedule and an external margin from the cost of the next module.

Figure 9 records all of the flows described in this paper as a supply matrix. Of the 272 possible directed pairs between seventeen verticals, 49 carry a described flow CALC, before counting research and development and carbon measurement, which relate to every vertical. Power from the Core reaches all sixteen other verticals CALC. Water from desalination reaches steel, concrete, molecules, data centers, housing and mineral refining. Steel reaches the machines, the Core works, the ports and housing. The matrix makes one property visible: most of its rows and columns intersect, so that an improvement in one vertical lowers costs in several others. The compounding described in Section 9 is physical before it is financial. Each year's output increases the system's ability to build the next year's capacity.

Figure 10 gathers, for the principal uses, what one tenth of one module could produce if it were dedicated entirely to each use in turn, expressed as a share of the corresponding world output. Each row is an alternative use of the same 244 TWh CALC, not a sum. The subsections that follow derive each value.

7.3 Green steel

The steel vertical converts Core power into hydrogen, hydrogen into direct reduced iron, and direct reduced iron into steel in electric arc furnaces. The route requires about 3.48 MWh of electricity per tonne of liquid steel EXT [23]. By Equation 10, one tenth of one module's output, 244 TWh CALC, could make 70.2 Mt of steel per year CALC, 3.8 percent of world crude steel output in 2025 CALC. The whole output of one module could make 702 Mt CALC, 38.0 percent of world output CALC. Making all of the world's 2025 crude steel by this route would require 6,436 TWh CALC, the output of 2.63 modules CALC. At the industry's average intensity of 1.92 tonnes of CO2 per tonne EXT [12], each 70.2 Mt made by this route in place of conventional steel avoids up to 135 Mt of CO2 CALC.

Inside the platform, steel is not primarily a product for sale. It is the skeleton of the platform: tunnel boring machines, turbine components, penstocks, marine works, HVDC structures, port equipment and housing. A project that depends on external steel markets inherits their prices, their delays, their geopolitical exposure and their stacked margins. Internalizing steel gives the platform control of its own schedule.

7.4 Green concrete

Concrete is the contact layer between the platform and the ground. It lines the subsurface works, supports the Core, builds ports and logistics infrastructure, and forms the foundations of housing and industry. The vertical uses Core power, platform water, aggregates recovered from excavation, supplementary cementitious materials, calcined clays and mineral fillers, under quality systems governed by the Charter; mix designs are withheld. The chemistry sets a limit that cheap power alone cannot cross: about 60 percent of the cement industry's CO2 is released by the calcination of limestone, not by fuel EXT [29, 28]. Clean power addresses the energy share. Reducing the process share requires less clinker in each tonne of binder, which is why this vertical rests on binder chemistry as much as on energy. Because energy, water, aggregates and binder chemistry are controlled inside one platform, concrete becomes cheaper, cleaner, traceable and available on the platform's schedule.

7.5 Green molecules

Hydrogen, ammonia, methanol and synthetic natural gas are the chemical base of industry. World hydrogen demand passed 100 Mt in 2025, but production of low-emissions hydrogen reached only about 1 Mt, under one percent of the total EXT [13]. Electrolysis at 52 kWh per kg of hydrogen (Table 2) turns one tenth of a module's output into 4.70 Mt of hydrogen per year CALC. That is 4.7 percent of world demand CALC and about 4.7 times the world's entire 2025 production of low-emissions hydrogen CALC. The whole output of one module would yield 47.0 Mt CALC, nearly half of world demand.

The decisive number is the cost of the electricity in each kilogram. At the posted tariff it is 1.30 US dollars per kg CALC. At the internal transfer price it is 4.2 US cents per kg CALC. At that price the green premium, the extra cost that climate policy usually asks markets to pay for clean molecules, ceases to be an energy cost. Hydrogen feeds steel and industrial heat; ammonia feeds fertilizer and maritime fuels; methanol feeds shipping and chemicals; and synthetic natural gas can use existing gas infrastructure, with its carbon inputs governed and measured. Ammonia is 17.8 percent hydrogen by mass CALC, so each tonne needs about 9.2 MWh for its hydrogen alone CALC. One tenth of a module could therefore supply the hydrogen for up to 26.5 Mt of ammonia per year CALC, about 14 percent CALC of the 185 Mt the IEA reports for 2020 EXT [31], and that ammonia would carry no gas price into fertilizer. The purpose is not a line of premium green products. It is to make clean molecules industrially ordinary.

7.6 Compute and chips

The compute vertical converts Core power into sovereign computing capacity, and the wafer factory gives the platform a path to controlled hardware provenance and reduced exposure to external semiconductor bottlenecks. One tenth of a module, 244 TWh per year CALC, is a continuous facility load of 27.9 GW CALC. That is 50.4 percent of all electricity used by the world's data centers in 2025 CALC, 25.9 percent of the IEA base case for 2030 CALC, and 1.39 times the entire US data center consumption of 2023 CALC. The advantage is not cheap electricity alone. It is firm power at the site, freedom from the interconnection queue (Section 3.4), cooling integrated with platform water, use of waste heat, secure siting and a controlled hardware path. For sovereign cloud customers, universities, regulated industries and states that do not wish to depend on a single foreign provider, the value is jurisdictional and physical control over the infrastructure that computes.

7.7 The machines that build the machine

Five verticals exist so that the platform does not wait for anyone else's factory: tunnel boring machines, turbines, HVDC equipment including transformers, the 3D factory for components and spares, and refined materials. Research and development serves all of them. The cost effect of this kind of integration has been measured by an incumbent's own method. In 2011 NASA estimated, with its NASA-Air Force Cost Model, that developing a launch vehicle like the Falcon 9 would have cost 4.0 billion US dollars under traditional contracting, or 1.7 billion under a more commercial approach; the vertically integrated developer reported, and NASA verified, about 390 million US dollars for the Falcon 1 and Falcon 9 together EXT [40]. The actual cost was 9.75 percent of the traditional estimate CALC.

The opposite base rate is equally documented: 8.5 percent of more than 16,000 large projects were delivered on budget and on time EXT [38], and large dams have cost on average 96 percent more than estimated EXT [39]. The causes are several; the research on large projects emphasizes optimistic estimates and long delivery windows as much as supply [38]. Dependence on external suppliers is the cause the platform is built to reduce. Each external seller adds a margin and a lead time, and under debt financing interest accrues while the project waits. A platform that makes its own machines, steel, concrete and power at cost has fewer parties to wait for and fewer margins to pay. It does not remove every dependency: specialized components, some materials, permits and skilled labor still come from outside, and the schedule depends on them.

7.8 Excavation, logistics, housing and ports

Excavation is not waste disposal; it is a supply. Rock removed from the subsurface works is characterized, sorted and routed to aggregates, binders, minerals or residual streams, each of which is measured and given a number in the public record. Logistics, housing and infrastructure carry goods and house the workforce. Transshipment and port operations give the platform its own quays for receiving equipment and shipping output. Steel and concrete from the platform build the quays, warehouses and housing, and the data centers optimize logistics scheduling.

7.9 Carbon measurement

The carbon vertical quantifies displacement, captures residual process emissions, and locks carbon into stable products and sinks. It produces the audited record of what the platform displaces, and it does not trade that record. CTMP will never sell carbon offsets or carbon credits TGT. The reason is structural as well as ethical. A platform designed to make clean energy cheaper than fossil energy has no need of a price on carbon, and an income from credits would give it a financial interest in the continued existence of the emissions it is built to end.

7.10 Pricing rules

Two pricing rules apply to every vertical. Inside the platform, goods and services move at computed cost with no margin TGT: steel to the turbine works, water to the concrete plant, power to the electrolyzers. Outside the platform, output is sold at about 20 percent below the prevailing market price TGT. The first rule prevents rent from hiding inside internal transfer prices, where it often accumulates long before it appears in public tariffs. The second places the platform's output in the market at a price that competitors must meet, which carries the lower cost floor beyond the platform's own customers.

8. Water as a solved input

Link 5, second part: when energy is no longer the binding cost, desalinated water becomes a public input rather than a premium commodity, and brine becomes a feedstock rather than a liability.

8.1 The energy content of water

Desalinated water is energy in liquid form. The minimum theoretical energy for seawater separation at 50 percent recovery is about 1.06 kWh per cubic meter EXT [26], and modern plants use 2.34 to 4 kWh per cubic meter in practice EXT [20, 21]. By Equation 8, the energy cost of one cubic meter at 3.0 kWh is 24.4 US cents at the 2024 US industrial average price CALC, 7.5 cents at the posted tariff CALC and 0.24 cents at the internal transfer price CALC (Figure 11, panel B). For comparison, the full levelized tariff of the Hassyan seawater reverse osmosis plant in Dubai, reported as a world record when its developer was selected in 2023, was 36.5 US cents per cubic meter, covering capital, membranes, operation and energy for 818,280 cubic meters per day EXT [41]. At the internal transfer price, the energy component equals 0.66 percent of that full tariff CALC. When energy is that cheap, the cost of water is set by intake, pretreatment, membranes, brine handling and distribution, not by power.

8.2 The host allocation

Each module allocates about 2 billion cubic meters of fresh water per year to its host country free of charge TGT [18], subject to verified distribution governance and operational constraints. That is 5.48 million cubic meters per day CALC, 5.7 percent of the world's entire desalinated water production in the most recent global inventory CALC [21]. At 3.0 kWh per cubic meter it requires 6.0 TWh per year CALC, 0.25 percent of the module's output CALC; across the range of specific energy found in practice, 2.34 to 4 kWh per cubic meter, it requires 0.19 to 0.33 percent CALC. Figure 11, panel A, divides the module's output into 400 equal parts: the host country's free water needs about one of them. The energy cost of the whole allocation is 4.8 million US dollars per year at the internal transfer price CALC, or 150 million at the posted tariff CALC. As a reference volume, 2 billion cubic meters per year supplies 100 liters per person per day to 54.8 million people CALC, or 50 liters per day to 109.6 million CALC. The allocation is not a gesture. It is a structural output of a platform whose energy cost no longer binds.

8.3 The energy ceiling

The host allocation is far below what the energy could supply. If one tenth of one module's output were dedicated to desalination at 3.0 kWh per cubic meter, it would produce 81.5 billion cubic meters per year CALC, 223 million cubic meters per day CALC, 2.34 times the world's entire desalinated water production in the most recent inventory CALC. At 100 liters per person per day, that volume would supply 2.23 billion people CALC, more than the 2.1 billion who lacked safely managed drinking water in 2024 EXT [9].

This calculation shows that energy is not the constraint on the world's drinking water. It does not show that the water would reach those people. Most of them live inland, in dispersed settlements, and many live in fragile states, where coverage of safely managed drinking water is 38 percentage points lower than elsewhere EXT [9]. Conveyance, storage, treatment at the point of use, governance and maintenance then become the binding constraints, and each has its own capital and energy cost. The platform's claim is narrower and exact: where seawater can be reached, the energy needed to make it drinkable is no longer scarce.

8.4 Brine as feedstock

Conventional desalination treats brine as a disposal liability. The world's plants discharge about 141.5 million cubic meters of it per day, 1.5 cubic meters for every cubic meter of fresh water EXT [21]. The platform treats brine as a feedstock. At a recovery ratio of 0.40, producing 2 billion cubic meters of fresh water draws 5.0 billion cubic meters of seawater and returns 3.0 billion cubic meters of brine per year CALC. Equation 9, with the reference composition of seawater [22], gives the mass of each major constituent contained in that feed: 99.2 Mt of chloride, 55.3 Mt of sodium, 13.9 Mt of sulfate, 6.58 Mt of magnesium, 2.11 Mt of calcium, 2.05 Mt of potassium and 0.345 Mt of bromine per year CALC.

Figure 12 compares three of these with world production in 2025 [14]. The magnesium contained in one year of feed is 5.98 times world primary magnesium metal production CALC; the bromine is 0.80 times world bromine production CALC; and the sodium chloride equivalent is 0.52 times world salt production CALC. These are contained masses, upper bounds on what can be recovered, not forecasts of recovery or sales. Their significance runs opposite to what an abundance argument might expect. For bromine and salt, one module's host allocation alone contains a quantity of the same order as the entire world market. The binding constraint on brine revenue is therefore market absorption, not supply. This is the quantitative basis of the rule, stated in the earlier edition of this paper, that brine economics are constrained by market share and not by production [1]. Recovery is sized to the platform's own use and to the share of each market that can be supplied without collapsing its price; the remainder is managed to reduce discharge. The platform states openly that recovery yield, processing cost and market value must be demonstrated at scale [18].

8.5 Allocation governance

Water is allocated as a public stability input. The host country takes what it needs free of charge, and surplus water goes to neighboring countries, also free of charge TGT. Distribution is verified, metered and reported in the same public record as power. The free water follows the build: the Core must exist before the water does, and the platform says so plainly [18].

9. Compounding

Link 6, first part: reinvestment and internal supply turn the platform into a system that builds more of itself.

9.1 The growth target

The platform targets growth in vertical output of 25 percent per year TGT. This edition replaces the 10 percent rate TGT of the May 2026 edition with that current target. Equation 7 gives the consequence. At 25 percent per year, output doubles every 3.11 years CALC and reaches 3.05 times its starting level after 5 years, 9.31 times after 10, 28.4 times after 15 and 86.7 times after 20 CALC. At 10 percent it doubled every 7.27 years and reached 6.73 times after 20 CALC (Figure 13). Over 20 years the two rates differ by a factor of 12.9 CALC. A target is not a measurement, and the rate actually achieved will be reported as it is measured.

9.2 Physical compounding

The growth is physical before it is financial. Steel made this year becomes next year's tunnel boring machines and turbines; concrete made from this year's excavation lines next year's tunnels; hydrogen made this year reduces next year's iron. Each internalized supply is designed to remove a delay and a margin from the next build, so that the cost and time of successive modules fall as the platform's internal capacity rises; the rate at which they fall will be measured and reported. Financial reinvestment (Section 5.7) supplies the capital; physical reinvestment supplies the materials and machines that the capital would otherwise have to buy. A project ends when it is built. A system that uses its own output to build more of itself does not.

9.3 The fleet

The deployment plan calls for at least 80 to 100 modules worldwide TGT, in each of which the Core's power and the desalination plant's water are shared with the remaining fifteen verticals in a circular arrangement TGT, with locations determined through The Wall and the review it opens (Section 11). At the design output, 80 modules would generate 195,523 TWh per year and 100 modules 244,404 TWh CALC, 6.2 to 7.7 times world electricity demand in 2025 CALC. Figure 16, panel B, sets the thresholds in module-years: 4.3 modules equal all of the world's coal-fired generation in 2025, 7.5 modules all of its fossil-fired generation, and 13.0 modules all of its electricity demand CALC. A fleet of the planned size is therefore not a contribution to the existing electricity system. It is an energy base several times larger than that system, intended for the uses that cheap firm energy makes possible: desalination at continental scale, electrified heavy industry, synthetic fuels and computation.

9.4 Bounds

Four constraints bound the rate at which the fleet can grow. The first is capital. The financing plan targets self-financing after the first two or three modules TGT; whether reinvested revenue covers both new Cores and the verticals they need depends on the combined cash flows described in Section 5.7, which are not modeled here. The second is absorption (Section 6): each module needs its load. The third is labor (Section 10): the workforce for construction and operation must be trained and housed. The fourth is verification (Sections 11 and 14): each site must pass public mandate, institutional readiness and physical review in a fixed order. Growth is governed by the slowest of the four, and it is reported as measured.

10. Workforce

Link 6, second part: the people who build and run the platform are one of its principal outputs.

10.1 Occupations

The workforce is a principal social output of the platform, not an input to be minimized. The platform requires tunnelers; marine, hydraulic and geotechnical engineers; metallurgists; turbine and HVDC specialists; semiconductor and data center technicians; concrete scientists; desalination operators; logistics and port personnel; geologists; quality auditors; safety professionals; and public governance staff. These are durable industrial occupations with careers measured in decades. Workers from coal, oil, gas, shipping, mining, construction and fabrication already carry many of the skills the platform needs. The transition the platform offers runs from extraction to the construction and operation of the energy layer, not from work to unemployment, and its scale creates training systems, apprenticeships, credential conversion and long employment ladders.

10.2 How the workforce is counted

The workforce is modeled vertical by vertical. A consolidated model aggregates seventeen vertical models, reports low, central and high corridors, and discloses two scopes, the industrial platform and the civic economy around it, at Year 1 and Year 30 [42]. One reporting rule is fixed. Recurring full-time-equivalent positions and construction job-years are reported separately and are never added into a single figure. They measure different things, people in continuing posts and person-years of finite construction work, and adding them would overstate employment. The Mobilization [42] sets out a first-year construction labor requirement larger than that of any project on record and identifies the national labor pool deep enough to meet it. Labor mobilization is accordingly listed in Section 16 as a binding constraint on the first module.

11. The Wall: demand recorded before allocation

Link 7, first part: where the platform is reviewed first is decided by public demand recorded under published rules, not by private selection. This section follows the procedure published in Order of Operations [2].

11.1 The order of operations

Large infrastructure has usually entered a country from above. Capital introduces itself to a ministry, a bank to a treasury, one government to another, and by the time the public hears of it the decisive conversations have already taken place in rooms the public was not in. The CTMP reverses that order. The first entry is not a delegation, a fund or a letter of intent. It is a voice: one anonymous entry on The Wall, recorded under a country's name, with no account, no name, no email and no payment. When enough of a country's people have made their voice heard, the count is complete, and only then does the procedure reach the state. The Stewardship Charter states the relationship in one line: "The Wall is not a petition to any government" [19]. What the count decides is narrow and exact: where there is enough human demand to justify looking seriously, and in what order. Engineering decides what is possible, governments decide what is lawful, and the physical world has the final word [2].

11.2 Definitions

11.3 Two levels with two purposes

The two counts are different thresholds with different purposes, and they are never conflated. The first count exists so that governments become familiar with the procedure by watching it operate, in public, on real countries before it reaches their own. The full count exists to open deployment. Reaching the first decides nothing about the second. Because both terms of each formula differ by the same factor, the full count is always exactly 100 times the first count CALC, at the floors as well: 2,500 is one hundredth of 250,000. For a large country, the first count is one person in two thousand CALC; India's is 720,000 TGT and its full count is 72 million CALC. For a country of one million people, the floor of 2,500 is one person in four hundred CALC. The floor asks more of a small nation as a share, so that a handful of people can never stand in for a country, and less of it in raw numbers, so that a small nation never has to outnumber a large one to be heard [2].

11.4 Ranking by completion

The race is ranked by completion, the share of each country's own first count that has been reached, not by raw totals. Table 9 shows the three leading countries at the time Order of Operations was written [2]. India's first count is 277 times Ireland's CALC, yet the two were separated by 0.15 percentage points of completion CALC. The ranking counts percentages; the finish line counts voices. Ireland, in third place, needed 1,355 more voices, while Papua New Guinea, in first place, needed 2,435 CALC, so the first country to complete its count need not be the one in first place. The order also moves. In the snapshot of 28 to 29 September 2026, India held 334,423 voices and stood third; it then added 11,439 and rose to second MEAS [2]. No one assigned that order, and no one can.

11.5 Three tests in series

A completed first count opens one thing: review. The country then faces three tests, in series. The Wall tests public will: did enough people move? Political integration tests institutional readiness: can the state respond? Technical verification tests physical reality: can the geography, geology, grid and corridors support a pathway? If any one of them fails, the country does not proceed, and its count stays on The Wall [2]. Part eight of review then examines the integrity of the result as a whole.

11.6 Eight parts in a fixed order

The order is not arbitrary. Part one comes first because nothing physical can be examined lawfully without a pathway. Part two confirms that the count was earned in public and not arranged in private. Parts three to seven question the ground, the sea, the rock, the land and the grid. Part eight comes last and asks the hardest question: whether the country can take part without corrupting the system that selected it. Parts two to eight are the review studies, which CTMP carries out in conjunction with the government at the depth the procedure sets for them; each asks whether the country is serious enough for deeper study, not whether a module can be built tomorrow [2].

11.7 Two clocks and seven statuses

Two public clocks govern First Independent Review TGT. The first runs seven days from the delivery of notice to the government. A response within seven days stops it, and the country is Responsive; if seven days pass without a response, the country is certified Non-responsive. The second runs thirty days from the government's response. Within it, the government supplies the information the studies require and CTMP completes the studies; if the information is not supplied, the country is Defaulted. A country that is Non-responsive, Declined or Defaulted keeps its count, but not the active position, which passes to the next eligible country to complete its first count, where the clocks start again. A country that is Not proceeding has failed a test; it also keeps its count, and the sequence moves on. Read together, the two clocks mean that no country can hold the active position for more than 37 days from notice without either proceeding or passing it on CALC [2]. The clocks bind CTMP as well. The completion of the count and the delivery of notice are both timestamped, so the record shows how quickly CTMP acted, and the studies are CTMP's to complete within the same thirty days.

There are seven statuses and no others: Pending, Responsive, Proceeding, Not proceeding, Non-responsive, Declined and Defaulted (Figure 15). A status records where a country stands, not what it deserves. The words are deliberately neutral, there is no score, and no motive is recorded, because no motive can be measured. Statuses are distinct from events. An event is something that happened, with a timestamp: first count completed, notice delivered, acknowledgment received, response received, redirected, information received, studies completed, status certified. Only two events move a clock: delivery of notice starts the first, and a response stops the first and starts the second. A response is a reply from a government office that accepts the notice, asks a question about it, names the proper authority or declines. An automatic receipt is an acknowledgment, not a response. A redirection to the proper authority is a response; it is logged as an event and restarts neither clock. A reply after the seventh day is logged as an event and does not change a certification already made [2].

11.8 What a government can and cannot do

A government is not the gatekeeper of the process: the count needed no permission to begin and does not pause because a minister is busy. A government is the gatekeeper of its own country's passage through the process: a state that declines has declined for itself, which is its right, and the record shows that it used it [2]. Notice is not permission. A veto would make the procedure depend on a government. Notice leaves a government exactly as sovereign as it was and makes what it does next visible.

11.9 Precedent and institutional memory

The first country to complete its first count opens Docket No. 1. Every government after it can read the dockets of those before it: the second reads one, the tenth reads nine. Which office took the file, how many of the seven days it used, what information was asked for within the thirty, and what happened next all become public precedent. A private interaction produces private memory; a public interaction produces precedent. The docket also keeps what institutions lose when ministers change, departments merge and files are archived. The measure of success is a sentence spoken inside a ministry that has not yet received notice: "We should know who handles this before our number gets there" [2]. That sentence describes familiarization: readiness chosen freely, before it is needed.

11.10 Integrity

The Wall asks for no identity. Limited technical controls prevent duplicate or abusive entries, but no account, name or email is required, and individual voices are not publicly identified [18]. The order of review cannot be bought with capital: the Charter states that donations "do not purchase influence" and that "the relationship between donation and CTMP access is zero" [19]. It cannot be bought with influence: "there are no exceptions sold for influence" [19]. It cannot be bought by a government, which is why part eight of review looks for queue-jumping, capture and off-book allocation. And it cannot be chosen by the author of the procedure, who has no means to select a first country [2]. On 29 September 2026 The Wall held 667,101 voices MEAS [2]; the live count is published at peoplesctmp.org.

11.11 The free variable

Every element of the procedure is determined: one hundredth, seven days, thirty days, thirty-seven at most, seven statuses, eight parts in a fixed order, and a docket with fixed fields. Given the same inputs, it does the same thing in any country, whoever holds office and whoever wrote it. One variable is not determined: whether anyone begins. No clause can produce a voice, no budget can buy one, no government can order one and no founder can summon one. The procedure sets two free wills side by side, the collective will of a people and the will of a state, and commands neither. Where they meet and the ground agrees, the country proceeds [2].

12. The Stewardship Charter: governance as part of the machine

Link 7, second part: a platform of this power can become an instrument of release from scarcity or an empire with better engineering. The Stewardship Charter determines which, by removing the surfaces through which infrastructure is usually captured.

12.1 The capture surfaces

Infrastructure is captured through a small number of surfaces, each of which converts control of an essential service into a claim on the people who use it. Table 12 lists them with the provision of the Charter that removes each one. The Charter is applied through an embedded enforcement layer of metering, telemetry, audit logic, allocation status, verified-output status and public reporting, designed to remove discretion from the places where capture normally enters.

12.2 Zero debt

No CTMP asset may be financed with debt, and no lender may hold a lien, security interest or equivalent encumbrance over platform assets, output or receivables. At this scale debt is not only a financing instrument. It is a control surface, because whoever holds the claim can direct the asset when payments are at risk. Section 5.4 measured what the rule removes from one module: between 75 and 273 billion US dollars of interest over the life of a loan, depending on rate and term CALC. Section 3.3 showed why the rule matters most where the platform is most needed: in economies where capital costs two to three times as much EXT, financing accounts for around half of the levelized cost of a new solar plant EXT.

12.3 No public-market listing

No CTMP vertical may be listed on a public stock exchange, and no synthetic public-market structure may represent ownership in one. A listing converts infrastructure into a trading instrument and subjects it to liquidity demands, speculation, index inclusion and quarterly pressure, none of which is aligned with a public function that runs for decades.

12.4 Capped external equity

External equity is capped. No single investor or coordinated group may control more than the permitted threshold in any vertical, beneficial ownership must be traceable through every layer, and a breach triggers a governance response rather than a negotiation. Capital can participate within these rules. The record also shows what happens to capital that does not accept them: offers have been made to the platform, and none was accepted [2].

12.5 Mandatory reinvestment

The platform exists to expand capacity, maintain resilience and lower scarcity pressure, so surplus cannot be withdrawn while its public function is underbuilt. The reinvestment shares of the financing plan, 50 to 60 percent of revenue for the first module and 80 to 90 percent for later modules TGT, are what turn the margin between cost and tariff into new modules (Section 5.7).

12.6 The posted tariff and the Verified resale ceiling

The Core sells at a posted tariff that no operator can raise, and power resold as CTMP Verified is capped at 5 cents per kWh all-in TGT. Output that cannot meet the verification conditions cannot be represented as CTMP Verified. This is not coercive price control; it is truth in labeling for a governed supply chain. If an intermediary adds an extreme margin, the record shows that the output is no longer delivered under Verified conditions.

12.7 Internal transfers at cost

Verticals transfer power, water, materials, gases, computation and services to one another at computed cost. Rent often hides inside internal transfer prices long before it appears in public tariffs, and a platform of seventeen verticals has 272 directed internal relationships CALC in which it could hide (Section 7.2). The Charter closes all of them with one rule.

12.8 Transparency by structure

Verification of every CTMP claim operates through four layers: the text of the Charter; aggregated public reporting; independent third-party attestation; and tamper-evident technical records. An output that cannot be verified through the required layers cannot be claimed as CTMP Verified. The system does not ask the public for trust. It creates a record, and the record is public: every correction is dated and displayed, mistakes included [18].

12.9 Influence, donations and carbon

Three further provisions close the remaining surfaces. There are no exceptions sold for influence. Donations do not purchase influence, and the relationship between donation and CTMP access is zero [19]. And CTMP will never sell carbon offsets or credits TGT, so that the platform never acquires a financial interest in the persistence of the emissions it displaces (Section 7.9).

12.10 The author is bound first

The Charter binds its own author before anyone else. The procedure that sets the order of review gives its author no means to choose the first country; if it did, the count would mean nothing, and every government would be right to treat it that way [2]. The first step of the procedure is left unwritten on purpose, so that no one, the author included, can start it on another's behalf. In the author's words: "A system its author could start, its author could also stop, steer or sell" [2]. Capital is welcome. Capture is not. Conventional infrastructure bends governance to suit capital; the CTMP reverses the sequence, so that capital must conform to governance.

13. Health and climate pathways

Consequence of the chain: firm clean power and abundant water act on the exposures that cause disease and on the emissions that drive climate change. This section quantifies the pathways without attributing outcomes that depend on models not yet published.

13.1 Exposure, not mortality

The humanitarian case is not separate from the engineering; it is the reason the engineering matters. Clean baseload power displaces combustion. Clean water interrupts pathways of disease. Lower industrial costs reduce the scarcity premium embedded in food, housing, transport, medicine and public services. Firm energy reduces the fragility of grids, and industrial employment restores agency. The exposures are counted: 655 million people without electricity, 2.1 billion without safely managed drinking water, 3.4 billion without safely managed sanitation, and 99 percent of the world's population breathing air that exceeds WHO guideline limits EXT [10, 9, 35]. This paper reports no estimate of deaths averted. The effect of any module on mortality depends on its location, on what it displaces, on how its water is distributed and on the quality of public health systems, and CTMP's own estimate is withheld until its methodology is published [18].

13.2 The electricity gap in module terms

If the output of one module were shared equally among the 655 million people without electricity in 2024 EXT, each would receive 3,731 kWh per year CALC, 96 percent of the world's average electricity demand per person in 2025, about 3,875 kWh CALC. The calculation is not a distribution plan; most of those people live far from any coast and far from each other. It establishes that the quantity of electricity needed to bring every unconnected person to the world average is of the order of one module. The constraint is delivery, which is the subject of the HVDC vertical and of the national grids that part seven of verification examines.

13.3 Displacement arithmetic

Equation 11 gives the gross emissions displaced when a module's output replaces existing generation. At the lifecycle median intensity of coal-fired generation, 820 g CO2-eq per kWh EXT [24], one module displaces 2.00 Gt CO2-eq per year CALC, 5.2 percent of world energy-related CO2 in 2025 CALC. At the median for gas combined cycle, 490 g EXT, it displaces 1.20 Gt CALC, 3.1 percent CALC. For comparison, the IEA estimates that solar PV, wind, nuclear power, electric cars and heat pumps together avoided around 3 Gt of CO2 in 2025 EXT [6]; one module displacing coal equals two thirds of that CALC. In generation terms, 4.3 modules equal all of the world's coal-fired electricity in 2025 and 7.5 modules all of its fossil-fired electricity CALC (Figure 16, panel B).

13.4 Displacement integrity

These figures are gross and conditional. They hold only where a module's output replaces existing combustion; output that adds to consumption without replacing anything displaces nothing. The fleet cannot displace more than exists: world energy-related emissions were 38.4 Gt in 2025 EXT, and no displacement claim can exceed that bound. Lifecycle intensities combine CO2 with other greenhouse gases and are compared here with energy-related CO2, so the shares are approximate. The platform's own lifecycle emissions, chiefly those embodied in the steel and concrete of construction, must be subtracted once they are measured; because the platform makes that steel and concrete with its own clean power and controlled binder chemistry, those emissions are themselves reduced by the platform. The carbon vertical (Section 7.9) exists to measure displacement integrity and to publish it. The climate value of the CTMP depends on replacing high-carbon production, not on expanding total consumption, and the Charter's reporting makes that distinction visible.

14. Verification gates

A canonical public paper does not pretend that execution variables do not exist. The CTMP is ambitious precisely because the remaining work is serious. Before deployment, each candidate site must pass the gates in Table 13, which map onto the eight parts of First Independent Review, onto the formal deployment process that follows the full count, and onto the proof ladder.

The list is not a weakness of the project. It is the difference between a serious infrastructure platform and a slogan. The public mandate procedure is live, the governance architecture is defined, and the next gate is evidence: country-specific verification, independent review, and a deployment path that survives contact with physical reality.

15. Discussion

15.1 What the arithmetic shows

Read as a whole, the arithmetic supports one conclusion about scale. Each of the platform's central quantities is of the same order as a world total. One module's output is 7.7 percent of world electricity demand CALC and larger than every national system except two CALC. One tenth of it could produce more than twice the world's desalinated water CALC, supply half of the world's data center electricity CALC, or produce several times the world's low-emissions hydrogen CALC. The brine from the host allocation alone contains a quantity of bromine comparable to world production. At this scale the design problem changes. It is no longer how to sell a product into an existing market; it is how to convert an energy base larger than most markets into goods without being captured by those who control the markets. That is why the argument cannot stop at the Core. Without absorption the output has no value (Section 6). Without integration the cost of conversion is set by others (Section 7). Without the Charter the margin becomes rent (Section 12). Without The Wall the decision about where to begin is made in private (Section 11).

15.2 Why integration and governance are inseparable

The two halves of the platform are usually discussed separately, as an engineering proposal and a governance proposal. The arithmetic shows why they are one. The cost advantage in Section 5 comes from the absence of interest and margins, and that absence is exactly what the Charter guarantees. If debt were admitted, the margin between the tariff and the cost of the Core would fall from a factor of 31 under the modeled Charter basis toward a factor of 5 under commercial financing CALC, and the difference would flow to lenders. If listing were admitted, the reinvestment shares would compete with distributions to shareholders. If internal transfer prices carried margins, the cost floor would rise at every one of the 272 internal relationships CALC. The governance is not an ethical addition to an engineering design. It is the mechanism that preserves the engineering result.

15.3 Relation to the existing transition

Nothing in this paper argues against solar, wind, storage, nuclear power or the expansion of grids. In 2025 those technologies met all of the growth in world electricity demand (Section 3.6), and they will continue to grow. The CTMP addresses a different variable. Variable generation adds energy where and when the weather allows; the Core supplies firm energy at one site, at a fixed price, and converts it into goods there. The two are complements. The Core offers a firm base on which variable sources can be integrated, and a source of the steel, concrete, HVDC equipment and computation that the wider transition itself requires.

15.4 What would change the conclusions

Two kinds of evidence will update the figures in this paper. Metered output from the 50 MW and 250 MW stages will replace design values with measurements, and the labels in this paper will change accordingly. The first docket will show how governments actually respond to a public procedure with fixed clocks, which no design document can predict. Both will be public.

16. Concessions

The following limitations are stated against the project's own interest.

1. Design values are not yet measurements. Every Core figure in this paper, including capacity, capacity factor, planning capital, internal transfer price and build time, is labeled Target or Modeled because it has not yet been metered, and every output and ratio derived from them is Calculated from those labels. The arithmetic of every section inherits them. The proof ladder exists to replace them, and until it has, every ratio built on them is a statement about the design.
2. The planning capital is not a contract price. The figure of 125 billion US dollars TGT is an owner-side planning target, before site-specific contingencies, and is subject to geology, bathymetry, local labor, procurement and independent review. On costs specified independently of the CTMP model, Section 5.3 shows that the posted tariff survives capital overruns of 3.7 to 8.6 times; the 31.3-fold figure of Case A holds only if the modeled cost relationship holds. Any overrun would still reduce the capital available for reinvestment and slow the fleet.
3. Absorption takes years. Almost no host grid can take a module's output. Until export lines and verticals are built, delivered output, revenue and reinvestment are bounded by absorption, and Section 6.4 shows that the ramp from partial to full absorption takes several years even at the target growth rate.
4. Self-financing is a target, not a result. The capital figures in Section 5.7 count power revenue only, at full external sale at the posted tariff. Internal use, partial absorption and host-country taxes reduce that revenue. The planning capital covers the Core only; the verticals carry their own construction and operating costs, and their revenues, which offset part of those costs, are not modeled here. Whether the platform finances itself after its first two or three modules depends on the combined cash flows of the Core and the verticals.
5. Firm output still needs reserves. Planned maintenance and forced outages require reserve capacity, and the concentration of 300 GW at one site makes the largest single loss a design question for the proof ladder and for independent review.
6. Brine markets are small. For bromine and salt, one module's feed contains a quantity of the same order as world production. Contained mass is not recoverable mass, and recoverable mass is not saleable mass. Recovery yield, processing cost and market value must be demonstrated at scale.
7. Water volume is not delivered water. The energy to desalinate water for every person who lacks it is available within one module, but conveyance, storage, treatment at the point of use and maintenance are separate infrastructures with their own costs. The platform's water is delivered where its distribution governance has been verified.
8. Displacement is conditional. The emissions figures in Section 13 are gross and apply only where combustion is replaced. Output that adds to consumption displaces nothing.
9. The procedure has not yet run. As of 1 October 2026 no country had completed its first count MEAS [2]. How governments respond to notice, to the clocks and to public statuses is unknown until the first docket is filled, and the first notice will probably need to be improved for its audience.
10. Labor is a binding constraint on the first module. The first year's construction labor requirement is larger than that of any project on record [42]. Mobilizing, training and housing that workforce is a gate in its own right.
11. Comparators have vintages and methods. The external statistics come from different years, 2011 to 2025, and from datasets with different definitions: generation and demand, lifecycle and energy-related emissions, generation cost and delivered price. Each comparison in this paper states its basis, but none is more precise than its least precise input.
12. Targets are not outcomes. The growth target of 25 percent per year TGT, the external price about 20 percent below market TGT and the fleet of 80 to 100 modules TGT are commitments. They will be reported as measured, and against these targets.

17. Conclusion: the decision layer

The Earth already supplies the forces this platform uses. Gravity, water, rock, minerals and human skill are not exotic. Hydroelectric engineering, desalination, steelmaking, concrete chemistry, electrolysis, high-voltage direct current transmission, data centers and semiconductor fabrication are established technologies. What has been missing is not a new technology. It is the integration of these technologies at the scale of the problem, under a governance structure strong enough to prevent capture. The maturity of the components does not by itself establish the performance of the integrated system. That performance is the design basis of this paper, labeled as such throughout, and the proof ladder is the means by which it becomes measurement.

The CTMP rests on four public pillars: a tariff-governed hydro Core designed for 300 GW of net firm baseload per module TGT; seventeen integrated verticals that turn power into water, materials, molecules, computation, employment and further capacity; the Stewardship Charter, which removes the principal capture surfaces and governs verified output; and The Wall, which makes public demand visible and sets the order of review through a published procedure. The arithmetic of this paper shows what the four pillars amount to together at their design values. By design, one module is a national-scale energy system at one site TGT. Its modeled cost sits far below its fixed price MOD TGT, and on costs specified independently of the model the price holds against capital overruns of 3.7 to 8.6 times the planning figure CALC. Its output, once converted, is of the same order as world production of the goods that energy makes CALC. Each of these statements becomes a measurement as the 50 MW and 250 MW stages report metered output. Its governance removes the debt, the listing, the resale margins and the internal tolls through which such power is usually turned into rent.

If the platform is implemented as designed, it lowers the physical cost floor of modern life and gives host nations a path to energy security, water security, industrial capacity and employment at a scale that ordinary development programs cannot reach. If it is implemented badly, with debt admitted, governance bent, resale opacity tolerated, internal tolls normalized or siting rushed, it becomes another captured megaproject. The engineering is powerful; the governance determines what the power becomes. That is why the architecture is strict.

The CTMP is not a business proposal. It is an attempt at stabilization, and its power is the collective will of the people who choose it. The Wall is live, the procedure is published, its clocks are set and its docket is formatted. What remains is the one variable the procedure does not determine: whether enough people, officials, engineers and institutions choose to test the platform against the physical world, under rules that cannot be bought.

Scarcity is a choice someone else made. I refused it.Christopher M. Coode

References

References are numbered in order of first citation. Statistics cited from news reports of official data identify both the report and the underlying source.

1. Coode C.M. The Engineered End of Scarcity: A Canonical Basic Public White Paper on the Canal and Transshipment Mega Project (CTMP). 2 May 2026. Superseded by this edition.
2. Coode C.M. Order of Operations. Protocol note, revision 2. People's CTMP; 1 October 2026. https://peoplesctmp.org/order-of-operations
3. International Energy Agency. Energy and AI. Paris: IEA; April 2025. https://www.iea.org/reports/energy-and-ai
4. Ember. Global Electricity Review 2026. London: Ember; 21 April 2026. https://ember-energy.org/latest-insights/global-electricity-review-2026/
5. Ember. Yearly Electricity Data: 2025 generation by country, region and source. https://ember-energy.org (accessed 2026). Country and regional totals as tabulated by Pew Research Center, 23 June 2026, https://www.pewresearch.org/?p=336187
6. International Energy Agency. Global Energy Review 2026. Paris: IEA; 2026. https://www.iea.org/reports/global-energy-review-2026
7. International Energy Agency. Key Questions on Energy and AI (2025 estimate of 485 TWh for global data center electricity use). Paris: IEA (accessed October 2026).
8. Shehabi A., et al. 2024 United States Data Center Energy Usage Report. Berkeley, CA: Lawrence Berkeley National Laboratory; December 2024.
9. WHO and UNICEF Joint Monitoring Programme. Progress on household drinking water, sanitation and hygiene 2000-2024: special focus on inequalities. Geneva and New York; August 2025. https://washdata.org
10. IEA, IRENA, UNSD, World Bank and WHO. Tracking SDG 7: The Energy Progress Report 2026. Washington, DC: World Bank; June 2026.
11. World Steel Association. December 2025 crude steel production and 2025 global crude steel production totals. Brussels: worldsteel; 23 January 2026.
12. World Steel Association. Sustainability Indicators 2025 and World Steel in Figures 2026. Brussels: worldsteel.
13. International Energy Agency. Global Hydrogen Review 2026. Paris: IEA; 2026.
14. U.S. Geological Survey. Mineral Commodity Summaries 2026: Magnesium Metal; Bromine; Salt. Reston, VA: USGS; February 2026. doi:10.3133/mcs2026
15. International Renewable Energy Agency. Renewable Power Generation Costs in 2024. Abu Dhabi: IRENA; July 2025.
16. Lazard. Levelized Cost of Energy+ (LCOE+), 18th edition. New York: Lazard; 16 June 2025.
17. U.S. Energy Information Administration. Electric Power Annual 2024, average revenue per kilowatthour by sector (released 16 October 2025).
18. People's CTMP. CTMP in Plain Words: Start Here. https://peoplesctmp.org/technical-record-simplified.html (accessed 16 October 2026).
19. People's CTMP. Stewardship Charter, Plain-Language Public Edition v3.2. 16 December 2025. Parts Three and Seventeen as quoted in Order of Operations.
20. Arab News. Saudi Arabia leads world with 16m cubic meters of desalinated water daily (Shuaiba Phase 5: 2.34 kWh per cubic meter overall). 2026. https://www.arabnews.com/node/2647115
21. Jones E., Qadir M., van Vliet M.T.H., Smakhtin V., Kang S. The state of desalination and brine production: a global outlook. Science of the Total Environment 2019; 657: 1343-1356.
22. Millero F.J., Feistel R., Wright D.G., McDougall T.J. The composition of Standard Seawater and the definition of the Reference-Composition Salinity Scale. Deep-Sea Research I 2008; 55: 50-72.
23. Vogl V., Åhman M., Nilsson L.J. Assessment of hydrogen direct reduction for fossil-free steelmaking. Journal of Cleaner Production 2018; 203: 736-745.
24. Schlömer S., et al. Annex III: Technology-specific cost and performance parameters. In: Climate Change 2014: Mitigation of Climate Change. Contribution of Working Group III to the Fifth Assessment Report of the IPCC. Cambridge University Press; 2014.
25. Hitachi Energy. The world's most powerful transmission system: Changji to Guquan UHVDC, ±1,100 kV, 12,000 MW, 3,293 km. 12 July 2020.
26. Elimelech M., Phillip W.A. The future of seawater desalination: energy, technology, and the environment. Science 2011; 333(6043): 712-717.
27. IEAGHG. Low-Carbon Ammonia Roadmap. Insight Paper 2023-IP03. Cheltenham: IEAGHG; March 2023.
28. Andrew R.M. Global CO2 emissions from cement production, 1928-2018. Earth System Science Data 2019; 11: 1675-1710.
29. Global Efficiency Intelligence. Global Cement Industry's GHG Emissions. 2021.
30. Eurostat. Euro area annual inflation up to 10.6%, October 2022. Euro indicators news release; 17 November 2022.
31. International Energy Agency. Ammonia Technology Roadmap. Paris: IEA; October 2021.
32. International Energy Agency. Cost of Capital Observatory. Paris: IEA; 2023 and 2025 updates.
33. World Bank. International Debt Report 2025. Washington, DC: World Bank; 3 December 2025.
34. Rand J., et al. Queued Up: 2025 Edition. Characteristics of Power Plants Seeking Transmission Interconnection as of the End of 2024. Berkeley, CA: Lawrence Berkeley National Laboratory; December 2025.
35. World Health Organization. WHO Air Quality Database, 2022 update. News release, 4 April 2022: 99 percent of the global population breathes air that exceeds WHO air quality limits.
36. International Hydropower Association, World Hydropower Outlook 2026, as reported in South China Morning Post, "Global hydropower outlook report puts China at the head of the pack", 2026; and Foreign Policy, "China's Tibetan Mega-Dam Is Veiled in Secrecy", 2025. Construction began 19 July 2025.
37. U.S. Energy Information Administration. Nuclear explained: U.S. nuclear industry (average capacity factor of 91 percent in 2025). Washington, DC: EIA; 2026.
38. Flyvbjerg B., Gardner D. How Big Things Get Done. New York: Currency; 2023.
39. Ansar A., Flyvbjerg B., Budzier A., Lunn D. Should we build more large dams? The actual costs of hydropower megaproject development. Energy Policy 2014; 69: 43-56. doi:10.1016/j.enpol.2013.10.069
40. NASA. Falcon 9 development cost estimate using the NASA-Air Force Cost Model (NAFCOM), 2011, as documented in NASA Technical Reports Server document 20170008895.
41. ACWA Power and Dubai Electricity and Water Authority. Hassyan Independent Water Producer: preferred bidder at a levelized water tariff of USD 0.36536 per cubic meter for 818,280 cubic meters per day. Reported by Arab News, 18 August 2023, and Energy and Utilities, 2023.
42. Coode C.M. The Mobilization. Workforce Series No. 2. People's CTMP; 2026. https://peoplesctmp.org/the-mobilization.html

Appendix A. Calculation register

Every derived value in the paper, with its basis. All results in this appendix are Calculated CALC; their inputs carry the labels given in the text and in Table 2. Values are computed at full precision and rounded for presentation.

Appendix B. Change log from the edition of 2 May 2026

The earlier edition [1] is superseded. Changes are of four kinds: corrections, updates of data or parameters, additions, and removals.

Appendix C. Glossary and symbols

Christopher M. Coode · Founder and CEO, Hutchison Lea ConneXions Inc. · Architect and Steward, People's CTMP · peoplesctmp.org