CALC-076

CAST's headline 6.6 cents restated as a levelised cost

Feeds CLM-0697. Answers QST-004. Basis: USD, BASIS YEAR NOT STATED BY THE SOURCE; real, discounted; 30-year life; 1 GW continuous.

Why it was needed. CAST's 6.6 US cents/kWh is the cheapest institutional SBSP figure in the corpus and is quoted as though it were an LCOE. It is not: it is an undiscounted ratio of total cost to total energy. Comparing it against Western LCOEs without restating it compares a figure that charges nothing for the cost of money against figures that do.

Inputs.

Working, transcribed.

Annuity factor for 30 years at rate r:   AF = (1 - (1+r)^-30) / r
     3 per cent   19.600
     7 per cent   12.409
    10 per cent    9.427

Annual energy at 1 GW continuous:  8.76 bn kWh

LCOE = 17,272.7 M USD / (8.76 bn kWh x AF)

     3 per cent   10.1 cents/kWh    = 1.53 x the 6.6 headline
     7 per cent   15.9 cents/kWh    = 2.42 x
    10 per cent   20.9 cents/kWh    = 3.18 x

Against NASA, at the 3 per cent restatement:
    0.61  / 0.101  =   6.1 x        NASA RD1 baseline
    1.59  / 0.101  =  15.8 x        NASA RD2 baseline
    0.20  / 0.101  =   2.0 x        NASA electric-propulsion figure

Result: 10.1 cents/kWh at 3 per cent real, rising to 20.9 at 10 per cent. NASA remains 2.0x to 15.8x more expensive depending which NASA case is used.

Conclusion, and it is a basis correction rather than a disagreement. The gap between CAST and NASA is not primarily about inputs - it survives restating CAST onto a Western discounting convention. What the restatement removes is the part of the gap that was an artefact of CAST not charging for the cost of money, which is 1.5x to 3.2x of it.

Where this fails. - The spend profile is crude and the claim says so. All capital at t=0 is the limiting case that maximises the discounting penalty. A realistic multi-year build would give a figure between 6.6 and 10.1 cents at 3 per cent. - CAST's basis year is unknown, so this mixes an unknown-year USD against NASA's FY2022 with no adjustment possible in either direction. - CAST's underlying cost inputs are not reproduced in its paper. This corrects a basis mismatch, not an input disagreement. - Must not be quoted as CAST's own figure. It is a restatement of it, made here. - VER-825 PARTIAL 2026-09-05.

A LABELLING DEFECT IN THE CLAIM, RECORDED NOT FIXED, AND IT IS THE SAME CLASS AGAIN. The claim field reads "Against NASA OTPS's RD1 baseline of 0.61-1.59 USD/kWh". 0.61 is RD1's figure and 1.59 is RD2's - CALC-054's transcription of the published table gives RD1 DCF LCOE 0.61 and RD2 1.59. A range labelled "RD1 baseline" that in fact spans RD1 to RD2 attributes RD2's cost to RD1. The two ratios that follow, 6.1x and 15.8x, are individually correct against their own designs, so no number moves; only the label on the range is wrong. Recorded because a reader would take 15.8x as an RD1 result.

A REPAIRED MISATTRIBUTION IN THE SUPERSEDED CLM-0690, CARRIED HERE BECAUSE IT IS A FAILURE CLASS AND NOT A TYPO. CLM-0690 attributed "0.207 USD/kWh" to NASA OTPS. SRC-012 says "about 0.20". 0.207 is a model output computed here, moved into NASA's column. Caught by an independent verifier the same evening it was written. The correction moves the derived ratio only from 2.06x to 1.99x, and that is precisely why it was easy to miss. A derived figure and a source figure must never sit in the same sentence without their origins attached, and this record states the origin of every number in its inputs table for that reason.

Cite this