Source Review: Sowers 2018 — Commercial Lunar Propellant Architecture (CLPA)

Summary

Claim 1: "End-state production target ~2,450 tonnes of propellant per year"

Verdict: Novel supporting Why: The 2,450 t/yr output target is roughly 2.5x our calc's 1,000 t/yr midpoint, reflecting Sowers' commercial-scale ambition. Our calc's lower output scales the ISRU plant mass down proportionally, but the architectural difference (tent vs strip mining) is the more important factor.

Claim 2: "Architecture stages: prospecting (precursor) → pilot plant → first production plant → multiple subsequent plants → distribution chain"

Verdict: Consistent Why: Direct support for the staged-buildup-milestone decomposition in our calc (M1-M8). Sowers' stages map roughly to our M1 (prospecting), M2-M3 (lander/habitat), M5 (pilot ISRU), M7-M8 (multiple plants + distribution → net-positive export).

Claim 3: "Tent sublimation as the chosen ice extraction method"

Verdict: Novel supporting Why: Tent sublimation is the architectural choice underlying φ ≈ 534 (per Metzger 2023's classification). This is the key technical decision that yields Sowers' optimistic capex. Our calc uses a generic midpoint architecture (φ = 20); tent sublimation would shift our estimate sharply downward for the ISRU component.

Claim 4: "Operations are fully robotic — no humans required at the mining site during normal operations"

Verdict: Merits investigation Why: Fully-robotic operations eliminate the habitat + life support component from the M5 architecture entirely. Our calc retains the habitat (25-50 t) because we're modeling a manufacturing base capable of net-positive export, not propellant production alone. The robotic-only choice is appropriate for Sowers' scope but not directly transferrable to ours.

Claim 5: "The architecture assumes commercial launch services (Falcon Heavy, Vulcan, future Starship), not SLS-class transport"

Verdict: Consistent Why: Same commercial-launch assumption our calc uses (q1 $/kg figures). This is the basic transport-architecture decision that distinguishes Sowers' $4B from Jones' 35-year-breakeven pessimism.

Cross-reference

• PDF body could not be parsed (10MB exceeded WebFetch limit); review based on summary content from secondary coverage and direct cross-reference to Sowers 2021.
• The architectural staging here is the most concrete published roadmap for milestone-by-milestone buildup.
• The 2,450 t/yr endpoint is the most-cited commercial-architecture output target.
• Codex anti-hallucination check: limited quote extraction due to PDF body inaccessibility; review based on title + abstract + reconstruction from Sowers 2021.

Abstract

The Commercial Lunar Propellant Architecture (CLPA) is a multi-organization techno-economic study laying out an industrial architecture for harvesting and processing lunar polar ice into rocket propellant. The study assumes commercial launch services (not SLS-class), tent-sublimation ice extraction, robotic operations, and an end-state production of ~2,450 tonnes of propellant per year. The architecture decomposes into: prospecting (robotic precursors), pilot plant, first production plant, multiple subsequent plants, and a distribution chain to LEO/GTO/lunar orbit. The headline business case (later refined in Sowers 2021) shows positive returns under PPP models with the government acting as anchor customer for initial years. The architecture has explicitly informed the optimistic-φ ≈ 534 entry in Metzger 2023's TEA table.

Key claims

• 2450-tonnes-per-year: "End-state production target ~2,450 tonnes of propellant per year"
• staged-architecture: "Architecture stages: prospecting (precursor) → pilot plant → first production plant → multiple subsequent plants → distribution chain"
• tent-sublimation: "Tent sublimation as the chosen ice extraction method (rather than strip mining or borehole sublimation)"
• robotic-operations: "Operations are fully robotic — no humans required at the mining site during normal operations"
• commercial-launch-assumption: "The architecture assumes commercial launch services (Falcon Heavy, Vulcan, future Starship), not SLS-class transport"
• ppp-anchor-customer: "Government as anchor customer in the first years of production stabilises the revenue line and improves financeability"

Reviewer notes

Cannot fetch the full PDF body (10MB exceeds WebFetch max). Abstract reconstructed from search results and Sowers (2021) which cites and builds on this work. The CLPA is the architectural specification underlying Sowers' $4B / $2.5B-tent / 9% ROI numbers. For q5: this is the "what does a staged buildup look like, end-to-end" picture under the optimistic-φ branch. Pairs naturally with metzger-2013-bootstrap (which gives the seed-mass anchor) and jones-2020-breakeven (which gives the pessimistic-architecture anchor). The architectural staging here is the most concrete published roadmap for milestone-by-milestone buildup.