Method, assumptions and derivations

A.1 Approach and conventions

The paper was prepared by Cameroon Steel from published sources consulted up to 10 October 2026, listed in Annex C, and from Cameroon Steel’s reference configuration RC-1 (edition 10 October 2026), a Class 5 planning basis cited as the Cameroon Steel reference configuration RC-1 (planning estimate). This edition uses the RC-1 values as revised on 10 October 2026: domestic scrap available to the plant of about 30,000 t a year (range 0 to 60,000); a Phase 1 nameplate of 300,000 t of finished products with an expected sustained output of 270,000 to 300,000 t; Phase 1 inbound port tonnage of about 351,000 t a year; and a schedule with works from 2029, a first heat in the second half of 2030 and ramp-up in 2031. Every number carries a unit, a year or date and a source marker. Figures carry one of four status chips: sourced (read directly from a published source), derived (arithmetic on sourced inputs, shown below), assumed (a stated planning assumption) or illustrative (a diagram or scenario, not a forecast).

Project figures from RC-1 also carry one of three qualifiers, used only for RC-1 values and never as figure statuses: planned (a value of the current configuration that Cameroon Steel intends to build, such as the Phase 1 units and nameplate), target (a date or performance that the plan aims for but that depends on approvals, financing or suppliers, such as the first heat) and indicative (a value for a later phase or option that is not committed and may change substantially, such as Phases 2 and 3 and the direct-reduction module).

Units are metric. "t" is the metric tonne; "Mt" a million tonnes; "Mtpa" million tonnes a year; "kt" thousand tonnes. Energy is in GJ, MWh, GWh or TWh; 1 MWh equals 3.6 GJ. Natural gas volumes follow the Cameroonian industry convention of standard cubic feet: 1 Mscf is a thousand cubic feet and 1 bcf a billion. For conversions this paper uses 1.094 GJ per Mscf (about 1.037 MMBtu per Mscf at 1.055 GJ per MMBtu), an assumption appropriate for pipeline-quality gas. Money is quoted in the currency and year of the source; FCFA is converted to US dollars only where stated, at the RC-1 planning rate of 600 FCFA per US dollar (range 560 to 650) [S03]. The paper gives no investment amounts for Cameroon Steel.

worldsteel and the IEA measure energy and CO₂ on different boundaries. worldsteel route averages include direct emissions, emissions from purchased electricity and upstream emissions of materials, and count electricity as primary energy; the IEA reports final energy and, for indirect emissions, uses a 2019 world-average grid factor of 538 g CO₂/kWh [S01, S02]. RC-1 intensities are expressed per tonne of finished product, not per tonne of crude steel, and use Cameroon’s grid factor. Comparisons between these bases are indicative.

A.2 RC-1 planning assumptions

Table A.1 RC-1 benchmark and planning assumptions used in the paper (planning value and range)

Source: Cameroon Steel reference configuration RC-1 (planning estimate), edition 10 October 2026 [S03], and benchmarks as cited.

A.3 Derivations

D1. Route capex ranges (Chapter 2, Figure ES.1, Figure 2.3). The IEA gives annualised capital costs of US$52 to 94 per tonne of output a year for the blast-furnace route, US$53 to 136 for gas DRI-EAF and US$34 to 58 for scrap-EAF, at an 8% discount rate, 25 years and a 90% capacity factor [S02]. The capital recovery factor at 8% over 25 years is 0.0937. Overnight capex per tonne of annual capacity equals the annualised cost multiplied by the capacity factor and divided by the recovery factor: for example 52 × 0.9 ÷ 0.0937 ≈ US$500. The resulting ranges are about US$500 to 900, 510 to 1,310 and 330 to 560 per tonne, in about 2019 dollars. They are not escalated to current prices; budgets announced between 2020 and 2026 (Figure 2.3), which are broader in scope, run from about US$660 to 1,560 per tonne. The ranges serve to compare routes, not to price a plant. Each project benchmark is the announced budget divided by the announced capacity. The three US announcements give capacity in short tons (1 short ton = 0.9072 t), which are converted to tonnes: US$300 million for 500,000 short tons, or 0.454 Mt, gives about US$660 per tonne [S58]; US$350 million for 430,000 short tons, or 0.390 Mt, about US$900 [S57]; and US$450 million for 500,000 short tons about US$990 [S59].

D2. CO₂ reduction of electric routes (Figure ES.1, the executive summary and Chapter 2). 1 minus 1.47 ÷ 2.34 = 37%; 1 minus 0.69 ÷ 2.34 = 71% [S01].

D3. Gross CBAM charge (Chapter 2). Route-average intensity multiplied by the Q3 2026 certificate price of €82.32 per tonne [S54]: 2.34 × 82.32 ≈ €193; 1.47 × 82.32 ≈ €121; 0.69 × 82.32 ≈ €57 per tonne, before deductions for EU free allocation and any carbon price paid in the country of origin. Because worldsteel averages include purchased electricity and upstream materials, which CBAM does not price for iron and steel, these are upper bounds, not estimates of an exporter’s charge; White paper 06 (Table 5.2) computes the charge on the CBAM boundary.

D4. Grid emissions of EAF electricity (Chapter 2). 450 kWh × 0.226 kg CO₂/kWh ≈ 0.10 t CO₂ per tonne of liquid steel [S03, S07].

D5. Domestic scrap collected (Chapter 3, Figure 3.2). Domestic production of about 260,000 t (2022) [S74] × about 1.21 t of metallic charge per tonne of product (1 ÷ (0.88 × 0.94), the RC-1 yields) × about 40% local scrap share [S32, S33] ≈ 125,000 t, reported as "of the order of 0.1 Mt". The result is an order of magnitude, sensitive to each input; White paper 04 derives about 108 kt from 2023 billet imports (range 80 to 135 kt).

D6. RC-1 balances (Chapters 3 and 9). Liquid steel = finished products ÷ 0.94; metallics = liquid steel ÷ 0.88; DRI/HBI = 30% of metallics; scrap = the balance. Electricity = liquid steel × (450 + 70) kWh + finished products × 110 kWh + Phase 1 products × 300 kWh for induction heating. Gas = Phase 2 and 3 products × 1.3 GJ. Scope 1 = liquid steel × 0.10 t + gas × 0.0561 t/GJ; scope 2 = electricity × 0.226 t/MWh; upstream of imported DRI = DRI × 0.7 t [S03]. For Phase 1: 300,000 ÷ 0.94 = 319,149 t liquid; ÷ 0.88 = 362,669 t metallics; 30% = 108,801 t DRI; electricity = 319,149 × 0.520 + 300,000 × 0.410 = 288,957 MWh (average 33 MW over 8,760 hours; RC-1 reports 36 MW on an 8,000-hour operating basis); scope 1 + 2 = (31,915 + 65,304) ÷ 300,000 = 0.324 t per tonne; including upstream DRI (76,161 t) = 0.578 t. RC-1 electricity balances use 450 kWh/t in the furnace. At the upper planning value of about 490 kWh/t (D9), electricity rises by 40 kWh per tonne of liquid steel: about 13 GWh in Phase 1 (302 GWh; 38 MW on the 8,000-hour basis), 21 GWh for Phases 1 and 2 (443 GWh; 55 MW) and 43 GWh at full development (796 GWh; 99.5 MW). Scope 1 + 2 intensity rises by 40 × 0.226 ÷ 0.94 ≈ 0.01 t per tonne of product, to about 0.33 t in Phase 1 and 0.34 t at full development, and to about 0.59 t including upstream DRI.

D7. Billet share of chapter 72 imports (Figure 3.1). 161,400 t of HS 7207 [S05] ÷ 304,062 t of chapter 72 imports in 2023 [S71] ≈ 53%. The two sources use different reporting systems, so the share is indicative.

D8. Tap-to-tap time and furnace hours (Chapter 5, Figure 5.2). Heat energy = about 40 t × 450 kWh/t = 18 MWh at the planning value, or about 40 t × 490 kWh/t = 19.6 MWh at the upper value. Active power: about 25 MW, assumed for a furnace transformer in the 30 to 33 MVA class at an average power factor of about 0.8 (24 to 26 MW). Power-on time is about 43 minutes at 18 MWh and about 47 minutes at 19.6 MWh. Power-off times as in the published benchmark design [S75], with one extra minute for charging a smaller bucket: 12 minutes. Tap-to-tap is about 55 or 59 minutes. Heats at nameplate = 320,000 t ÷ about 40 t = about 8,000. Furnace hours = about 8,000 × tap-to-tap time ÷ 60: 8,000 × 55 ÷ 60 ≈ 7,300 at the planning value and 8,000 × 59 ÷ 60 ≈ 7,900 at the upper value, out of the 8,760 hours in a year. Specific transformer power of the benchmark design = 156 MVA ÷ 150 t ≈ 1.04 MVA per tonne [S75]; the Phase 1 furnace’s is lower.

D9. EAF electricity and DRI share (Chapter 5, Figure 5.3). Electricity = 450 kWh/t, read as an all-scrap value, + 14.5 kWh/t for each 10 percentage points of DRI in the charge, the ENERGIRON reference [S20]: 493.5 kWh/t at 30%, used as the upper planning value of about 490 kWh/t, and 551.5 kWh/t, about 550, at 70%. RC-1’s planning value of 450 kWh/t lies within the IFC range of 347 to 500 kWh/t, which already spans common charges; the penalty is applied only to derive the upper value.

D10. Gas demand of a DRI module (Chapter 4, Figure 4.4). Gas = DRI tonnage × 10 GJ/t [S20]; 1 Mt DRI → 10 PJ ÷ 1.094 GJ per Mscf ≈ 9.14 million Mscf ≈ 9.1 bcf; 0.5 Mt → 4.6 bcf. Share of 2025 national production (75.475 bcf) [S21]: 9.14 ÷ 75.475 ≈ 12%. LNG feed in 2025 = 17.905 + 53.716 = 71.6 bcf ≈ 95% of output [S21].

D11. Gas cost per tonne of DRI (Chapter 4). 9.14 Mscf per tonne × 1,830 FCFA (power-plant price) ≈ 16,700 FCFA; × 3,135 FCFA (industrial price) ≈ 28,700 FCFA [S21]; at 600 FCFA per US dollar, about US$28 to 48 per tonne. Chapter 4 quotes the range of White paper 05, Chapter 7: US$27 to 51.

D12. Ore, concentrate and pellets per tonne of DRI (Chapter 4, Figure 4.3). Pellets per tonne of DRI ≈ 1.45, implied by the BloombergNEF figures cited by IEEFA (210 Mt of pellet supply capping DRI at about 145 Mt in 2030; 252 Mt capping DRI at about 174 Mt in 2040) [S17]. Concentrate ≈ pellets (binder and moisture ignored). Ore = concentrate ÷ 0.45 mass yield [S19] ≈ 3.2 t per tonne of DRI. For 1 Mt of DRI: about 1.45 Mt of pellets and 3.2 Mt of itabirite a year.

D13. Electricity for hydrogen-based DRI (Chapter 4). 54 kg of hydrogen per tonne of DRI [S85] × 50 MWh per tonne of hydrogen (assumed) = 2.7 MWh per tonne of DRI; for 1 Mt, 2.7 TWh a year, against national generation of 8.12 TWh in 2025 [S51], about 33%.

D14. Reheating CO₂ (Chapter 6, Figure 6.2). Induction: 300 kWh × 0.226 kg/kWh ≈ 0.068 t CO₂ per tonne. Gas: 1.3 GJ × 0.0561 t/GJ ≈ 0.073 t CO₂ per tonne [S03, S07, S96]. Hot charge at 700 °C: 1.1 to 1.6 GJ minus 0.55 GJ [S50] = 0.55 to 1.05 GJ per tonne.

D15. Energy-management saving (Chapter 8). 4.1% first-year improvement [S31] × 289 GWh [S03] ≈ 12 GWh a year; presented as "about 4%".

D16. Full plant with domestic DRI (Chapter 9). At 1 Mtpa with 70% DRI in the charge: DRI = 0.70 × 1,208,897 ≈ 846,000 t. Shaft-furnace emissions from gas = 846,000 × 10 GJ × 0.0561 ≈ 475,000 t [S20, S96]; its electricity is counted below, with the plant’s. EAF direct = 1,063,830 × 0.10 ≈ 106,000 t. Reheating gas = 910,000 GJ × 0.0561 ≈ 51,000 t. Electricity = 753 GWh (RC-1) + 1,063,830 × 58 kWh (four further steps of 14.5 kWh/t from 30% to 70% DRI) ≈ 62 GWh + 846,000 × 73 kWh for the shaft furnace ≈ 62 GWh ≈ 877 GWh × 0.226 kg [S07] ≈ 198,000 t. Total ≈ 830,000 t ÷ 1,000,000 t of product ≈ 0.83, reported as "about 0.8 t CO₂ per tonne", excluding pellet production and mining.

D17. Iron units relative to domestic scrap (Chapter 3, findings). 1,208,897 t ÷ about 125,000 t (D5) ≈ 9.7, reported as "roughly ten times". Phase 1: 362,669 ÷ 125,000 ≈ 2.9, or about 3.6 against the rounded order of magnitude of 0.1 Mt, reported as "between three and four times"; 362,669 ÷ 53,000 to 61,833 ≈ 5.9 to 6.8 ("about six times").

D18. Phase 1 output range and transformer rating (Chapter 5, Finding 09). Heats = operating hours × 60 ÷ tap-to-tap time (D8). At 7,000 to 7,700 operating hours: about 7,640 to 8,400 heats at about 55 minutes, or about 305,000 to 336,000 t of liquid steel; about 7,120 to 7,830 heats at about 59 minutes, or about 285,000 to 313,000 t. Finished products = liquid steel × 0.94, capped at the mill’s nameplate of 300,000 t: about 287,000 to 300,000 t at 450 kWh/t and about 268,000 to 294,000 t at 490 kWh/t, reported as about 270,000 to 300,000 t, consistent with RC-1’s expected sustained output [S03]. Transformer for 320,000 t of liquid steel in 7,000 hours at 490 kWh/t: tap-to-tap = 7,000 × 60 ÷ about 8,000 ≈ 52.5 minutes; power-on = 52.5 minus 12 = 40.5 minutes; active power = 19.6 MWh ÷ 0.675 h ≈ 29 MW; ÷ about 0.8 ≈ 36 MVA, reported as about 36 to 40 MVA. Curtailment: at 55 to 59 minutes per heat, one hour of lost furnace time costs roughly one heat of about 40 t.

D19. Port inbound tonnage (Chapters 3 and 9). Inbound = metallics minus domestic scrap (30,000 t) plus about 6% of finished products for consumables (lime, alloys, electrodes, refractories) [S03]: 362,669 minus 30,000 + 18,000 ≈ 351,000 t in Phase 1; 1,208,897 minus 30,000 + 60,000 ≈ 1,239,000 t at full development, before any domestic DRI. Imported scrap in Phase 1 = 253,868 minus 30,000 ≈ 224,000 t.

D20. Iron units per tonne of billet (Chapter 2). 1 ÷ 0.88 metallic yield ≈ 1.14 t of iron units per tonne of liquid steel [S03]; with casting losses of about 1% (assumed), about 1.15 t per tonne of billet.

A.4 Limits of the analysis

The paper uses published benchmarks and a Class 5 planning configuration. Supplier-specific designs, site conditions and contract terms will change many values, and some inputs rest on dated or press sources (for example the 2006 scrap data and the industry’s reported scrap share). Capex values are benchmarks for comparing routes; they are not estimates of the cost of Cameroon Steel’s plant. Emissions estimates for future phases depend on the grid’s evolution, on hydrogen blending and on the boundary chosen; they will be replaced by measured values once the plant operates.