READING 01IEA · published Jan 2025
48 / 52
Construction starts · 2017–2024
Of 52 nuclear reactor construction starts worldwide, 48 used Chinese or Russian designs.
Civilian nuclear power · Industrial deployment · Evidence cut 15 August 2026
Siderian
How to Win the Second Nuclear Race
There is a nuclear race, and the West is losing.
The race is to build civilian nuclear power plants as repeat fleets: one proven design, supplied and constructed again and again.
Siderian is building the decision and evidence system that identifies which Western AP1000 projects are ready, verifies the suppliers behind them, protects one standard design, and assigns scarce industrial capacity by evidence.
One fleet. One reference design. One evidence trail. Positions earned by readiness.
The case in three readings
The race is industrial, immediate, and economic.
Starts preserve factories, skilled teams, quality records, and financing confidence. Announcements preserve none of them.
Evidence cut · 15 August 2026
READING 02IEA · published Apr 2025
415 → 945
Data-centre electricity · TWh
Data-centre electricity, 2024 → 2030, IEA Base Case.
READING 03IEA/NEA + INL · assumptions disclosed
$50–60
Economic objective · USD/MWh
The target range for repeat Western nuclear deployment. An objective, not a forecast or guarantee.
Observed records, published forecasts, and model outputs are labeled separately.
Open evidence ledger →Part I · Industrial diagnosis
How the West lost the first industrial nuclear race
The West did not lose nuclear because nuclear failed. It lost because it stopped building long enough for the industrial base to forget how.
The loss was not operating capability — the United States still runs the world's largest nuclear fleet. The loss was serial new-build capability: the institutional habit of turning the same design, the same suppliers, the same records, and the same construction discipline into repeat deployment.
France proved that scale works when continuity is protected. China and Russia then preserved the deployment machine while Western demand fragmented. The West still has vendors, utilities, regulators, and capital. It has pieces, not a system.
Figure 1 · The West built first. China built cadence.
Sourced public-data figure
Operable reactors, five-year intervals plus 2026 year-to-date. The United States built early, then plateaued. Germany built, then exited. China started late and reached 64 operable units. Source: World Nuclear Association reactor register and country profiles, cut 15 August 2026; historical year-end counts reconstructed from grid-connection and closure records.
World Nuclear Association reactor register and country profiles. Full reconstruction method in the evidence ledger.
Figure 2 · The 52-position construction registry
Sourced public-data figure
01
02
03
04
05
06
07
08
09
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
- 25
- Chinese design
- 23
- Russian design
- 2
- UK / European designs
- 2
- South Korea / national
Each square is one reactor construction start. China and Russia account for 48 positions; UK or European and South Korean programs account for four.
IEA, The Path to a New Era for Nuclear Energy, January 2025.
The West has pieces. China and Russia have systems.
Repetition preserves capability. Stop-start procurement forces every project to rebuild the supply chain, records, workforce, and confidence that the previous project lost.
The industrial asymmetry
First race · finding 01
China builds a fleet, the West makes collectors' items.
The difference is not ambition. It is whether one design, one supply chain, one evidence base, and one construction sequence are allowed to repeat.
Demian Reidel · Founder's thesis
Part II · The live race
The West is losing the second industrial nuclear race
The first race was lost when demand became too thin to preserve the build machine. The second can be lost for the opposite reason: demand is arriving before the system is ready.
The score establishes the industrial gap. The second-race problem begins after demand appears: announcements can claim scarce production attention before licensing, finance, reference discipline, and supplier capacity are ready.
Strict reactor-building register
Every counted start. No announcements.
Cut · 15 August 2026
- 01Chinese
Lufeng 1
China · 24 Feb 25
- 02Russian
Leningrad 2-4
Russia · 20 Mar 25
- 03Chinese
Shidaowan 2
China · 08 May 25
- 04South Korean
Shin-Hanul 3
South Korea · 20 May 25
- 05Chinese
Taipingling 3
China · 10 Jun 25
- 06Chinese
Jinqimen 1
China · 10 Aug 25
- 07Chinese
Zhaoyuan 1
China · 18 Nov 25
- 08Chinese
San'ao 3
China · 19 Nov 25
- 09Chinese
Ningde 6
China · 16 Dec 25
- 10Chinese
Lufeng 2
China · 22 Dec 25
- 11Chinese
Bailong 1
China · 25 Dec 25
- 12Chinese
Xuwei 1
China · 16 Jan 26
- 13Russian
Kursk 2-3
Russia · 31 Jan 26
- 14Russian
Paks II-1
Hungary · 05 Feb 26
- 15Indian
Kaiga 5
India · 01 Mar 26
- 16Indian
Kaiga 6
India · 01 Mar 26
- 17Chinese
Jinqimen 2
China · 04 Apr 26
- 18Chinese
Taipingling 4
China · 10 May 26
- 19South Korean
Shin-Hanul 4
South Korea · 29 May 26
- 20Western
Darlington SMR-1
Canada · 26 Jun 26
15
China + Russia
1
West · Canada
4
Korea + India
Excluded: Uzbekistan's June RITM-200N ceremony. WNA does not classify the pour as safety-related reactor-building first concrete and does not list the unit under construction as of this cut.
Why now
Three clocks are aligning.
Demand and policy move faster than qualified industrial capacity.
- 01
Demand
IEA, Energy and AI
415 → 945 TWh
AI data-centre load, 2024 → 2030, on 2–3-year build cycles nuclear supply chains cannot match.
- 02
Finance
World Bank Group / IAEA, June 2025
World Bank + IAEA
The World Bank's first nuclear re-engagement in decades opens the door for other development banks and private investors.
- 03
Supply
U.S. DOE, June 2026
$17.5 billion
DOE conditional long-lead financing for 10 large-scale reactors, with utility partners still to be selected.
The economic objective
$50–60
USD / MWh
The West must make the target credible.
China has built the continuity, supply chain, and financing conditions that put new nuclear near this range in public models. The objective is to create an equally repeatable Western path.
Siderian does not set tariffs, finance plants, design reactors, execute construction, or guarantee LCOE. It is being built to improve the industrial evidence, sequencing, repetition, and risk allocation on which lower costs depend.
Figure 3 · One axis, explicit assumptions
Sourced public-data figure
The copper band is Siderian's $50–60/MWh objective. The points are published model outputs, not observed fleet averages. Financing assumptions change the result before the reactor changes.
IEA/OECD NEA, Projected Costs of Generating Electricity 2020, Table 3.13a; China's National Development and Reform Commission benchmark tariff; Idaho National Laboratory, INL/RPT-25-84701. All values and dollar years are disclosed in the evidence ledger.
Observed policy
CNY 0.43/kWh
China's 2013 benchmark on-grid tariff. It is a revenue rule, not plant cost or LCOE.
Public model
$49.92/MWh
China new nuclear at a 3% discount rate, 2018 USD and 85% capacity factor.
Western pathway model
$49 / $65
AP1000 tenth series with / without a 40% investment tax credit, 2024 USD. Preliminary, academic, not investor-grade.
The cost stack
Four variables. Three governable interfaces.
- 01
Build cost
Repeat design, qualified suppliers, workforce continuity, and controlled rework determine overnight capital cost.
Qualified orderbook, capacity ledger, reference-package control, and evidence reuse.
- 02
Build time
Longer construction adds site overhead and financing cost before a unit produces electricity.
Slot sequencing, long-lead verification, readiness gates, and continuity rules.
- 03
Cost of capital
Nuclear LCOE is highly sensitive to the discount rate because most cost is committed before operation.
Decision-grade evidence and an explicit capital-recognition standard.
- 04
Fleet performance
Capacity factor, operating cost, fuel, and asset life determine how capital cost is spread over output.
Outside Siderian's operating authority; retained as an explicit dependency rather than a claimed result.
The missing institution
A platform cannot win the race. A fleet can.
AP1000 is the nuclear power plant platform. Siderian is developing the decision and evidence system that could turn separate projects into repeat deployment.
Without a governing layer, demand fragments, suppliers cannot plan against a credible orderbook, design deltas destroy evidence reuse, scarce slots go to the loudest buyer, and capital prices every unit as another exception.
Input
Separate projects
Unequal readiness · competing claims · bespoke deltas
- 01
Qualify
Is the buyer and site ready enough to enter the fleet path?
- 02
Verify
Is supplier throughput real, evidenced, and available in the claimed window?
- 03
Protect
Does the project preserve the reference package and reusable evidence?
- 04
Govern
Should the position advance, cure, re-sequence, transfer, or release?
Output
One governed fleet
Positions earned by evidence
What Siderian proposes to do
Siderian proposes to qualify demand, verify capacity evidence, govern readiness gates, sequence scarce industrial slots, and preserve a decision record.
What it does not do
Siderian does not design, license, construct, finance, insure, own, or operate a nuclear power plant. It does not replace the NRC, another national regulator, the reactor vendor, the licensee, the EPC, the owner, or their legal duties.
Current status: proposed standard and founding-stage operating model. Independent assurance governance is in formation.
Read the proposed standard →The allocation · decision demonstration
Three plausible projects. One scarce position.
Choose which project should receive the position. Then apply the published evidence gates. The system may overrule prominence, political weight, or requested date.
Fictional, anonymized demonstration. It is not a customer qualification or assurance decision.
Capacity positionSIS-01
Claimants03
Evidence gates05
StateUnallocated
| Published evidence gate | Dossier A | Dossier B | Dossier C |
|---|---|---|---|
| 01Site basisIs the site and licensing path evidenced? | Conditional | Verified | Verified |
| 02Capital authorityIs there decision authority and credible funding evidence? | Verified | Verified | Conditional |
| 03Reference disciplineDoes the project protect the standard plant and control deltas? | Conditional | Verified | Missing |
| 04Capacity basisAre critical supply windows verified rather than named? | Missing | Conditional | Conditional |
| 05Evidence threadCan the material records survive transfer, delay, or supplier change? | Missing | Verified | Conditional |
Verified means supported by a decision-grade record. Conditional means material evidence remains open. Missing means the dossier cannot support reliance on that claim.
Inspect all gates →The standard behind the decision
Six laws for repeat deployment.
Readiness must be visible. The reference plant must remain controlled. One weak project must not be allowed to damage the fleet sequence.
- 01
Demand aggregates
Fragmented interest becomes a same-design orderbook suppliers can plan around.
- 02
Design freezes
The reference package is protected; deltas must earn their way into the fleet.
- 03
Capacity reserves
Verified throughput is reserved by readiness, collateralized by escrowed deposits.
- 04
Evidence compounds
Every claim is linked to its record, reusable across units and over time.
- 05
Capital recognizes
Output counts only when a counterparty that bears risk changes its terms.
- 06
Failure isolates
Weak units cure, re-sequence, transfer, or release before local failure becomes fleet failure.
Institutional test
Independence must be built, not claimed.
Siderian is designed for independent operation. That independence is not yet an accomplished fact: its assurance governance is being formed and no independent assurance authority has been appointed.
Assurance authority
in formation
Mandate and appointment criteria are being developed. No authority is appointed.
Independent technical review
in formation
Capability requirements are defined at a role level; appointments are pending.
Conflicts and recusal
proposed
Public policy outline published; operating procedure and register remain to be adopted.
Commercial separation
proposed
Target operating model published; legal and committee implementation remain in formation.
Founder's letter
There is a nuclear race, and the West is losing.
Demian Reidel
Founder, Siderian
The AI buildout is accelerating electricity demand. The IEA estimates that data centres used about 415 terawatt-hours in 2024 and could use about 945 by 2030. The world needs far more power that is clean, firm, and scalable. Nuclear is one of the few proven sources capable of delivering all three at system scale.
Look at the construction record. Since January 2025, twenty nuclear reactors have begun construction. Twelve use Chinese designs. Three use Russian designs. Two are South Korean, two are Indian, and one is Western: Canada's Darlington BWRX-300.
15:1
Against China and Russia, the score is now fifteen to one.
One Western start is not a fleet.
The countries that build repeat nuclear programs will shape the standards, supply chains, skills, and financing architecture of the next industrial era. That position is being formed now.
The West is not losing for lack of reactor technology. Western institutions helped pioneer it. The West is losing because it allowed serial deployment to disappear.
China builds a fleet, the West makes collectors' items.
No durable nuclear fleet built itself. France made a national program. South Korea built a standardized export system. China is sustaining the world's largest active construction program. Fleets are designed, financed, governed, and repeated.
My own record crosses physics, capital markets, central banking, public strategy, and board-level nuclear governance. It does not include operating a nuclear plant or delivering an AP1000, and Siderian will not pretend otherwise. That boundary is why independent technical judgment and assurance governance must be designed into the institution.
To win the second race, the West must turn credible reactor technology into a repeat fleet. That requires a system that qualifies demand, verifies capacity, protects the reference design, governs scarce production slots, and gives capital evidence it can use.
The economic objective is equally explicit: move repeat Western nuclear toward $50–60 per megawatt-hour. That is a target, not a promise. It is the standard against which the industrial system should be judged.
Founder · documented record
Demian Reidel
Demian Reidel is a physicist and economist whose career spans capital markets, central banking, public strategy, and nuclear governance. He founded Siderian to build the rules and evidence layer required for repeat nuclear deployment.
His role is systems architecture. Technical qualification, operating judgment, and assurance decisions must be independently governed.
- Scientific training
- Physics, financial mathematics, economics
- Institutional record
- Markets, central banking, public strategy
- Nuclear experience
- Board governance, not operations or AP1000 delivery
Private diligence
Request investor access.
A private diligence path for qualified counterparties. Confidential material is shared under NDA, on request.
No transaction terms are published on this site.