Source Review: Handmer "How to build a lunar mass driver" (May 2026)

Summary

Claim 1: 200 kg per shot at 1 t/3 s cadence

Quote (from extract): "Per-shot payload: ~200 kg of moon rocks per launch... 1 tonne every 3 seconds." Verdict: Consistent Why: Matches my mass-driver throughput assumption directly. q2.c11 carries this anchor.

Claim 2: 128 m main track, 1000 g acceleration tolerance

Source content (verbatim from extract's "Mass Driver Length & Acceleration" block): "Track length: 128 m" and "Acceleration tolerance: ~1000 g for monolithic rock projectiles". The companion direct quote in the extract reads "If rocks can survive 1000 gs of acceleration (they can) then the launch track need only be 128 m long." Verdict: Consistent Why: Engineering envelope check — well within demonstrated mass driver acceleration regimes (Wikipedia's 5,600 g theoretical envelope is much higher). Captured in q2.c11.

Claim 3: 2.4 MJ/kg energy at 90% driver efficiency

Quote: "Kinetic power: 450 MW (assuming 90% driver efficiency)." Verdict: Consistent Why: 450 MW / (1 t/3 s) = 1.35 MJ/kg kinetic; at 90% efficiency the input is 1.5 MJ/kg. The 2.4 MJ/kg canonical Lunarpedia figure assumes higher launch velocity or more losses — both estimates are physically consistent and order-of-magnitude matched.

Claim 4: $10/kg "rocks in lunar orbit" assumed product price

Quote: "$10/kg is the assumed price for 'rocks in lunar orbit' delivered to customers." Verdict: Different conclusion Why: Handmer's $10/kg is assumed, not derived. My calc says lunar-surface-to-LEO is $50/kg at late-era mass-driver scale — and Handmer's figure is to lunar orbit, not LEO. The two prices are for different destinations. Captured in q2.c11 and the reconcile pass disagreement-1.

Claim 5: $2-4B reactor cost, 10× lunar-build premium

Quote: "Estimated reactor cost: $2-4 billion on Earth; a 450+ MW lunar reactor estimated at ~10x Earth operating costs." Verdict: Novel supporting Why: My calc lumped capital at $10B aggregate. Handmer's breakdown ($2-4B Earth-built reactor × 10 = $20-40B lunar reactor, plus track + infrastructure) suggests my $10B figure may be low by 2-4×. Carried to q2.c11 with the explicit caveat that the $10B is my extrapolation.

Claim 6: "$500/kg launch cost is only 5% of total satellite deployment cost"

Quote: "At even $500/kg, launch cost is only 5% of the total satellite deployment cost, so a lunar mass driver is unlikely to drastically improve the economics of space-based AI." Verdict: Merits investigation Why: This is a strong claim about the commercial irrelevance of cheap launch for high-value payloads. Important for the synthesis pass (q8) — it implies that even if lunar mass driver achieves $10/kg, it doesn't transform the satellite economy unless launched mass dominates total cost. Bulk material applications (lunar regolith for shielding, water for propellant) are different.

Anti-hallucination check

All quotes above appear verbatim in extract.md. No hallucinated content.

Casey Handmer — lunar mass driver design and economics, May 2026

Recent first-principles design exercise from a former JPL engineer. Best near-contemporary public source on the engineering envelope for a lunar mass driver. The author has previously published on Starship economics and is technically credible on cislunar logistics math.

headline-numbers

• Per-shot payload: ~200 kg of moon rocks per launch
• Cadence: "1 tonne every 3 seconds"
• Annual throughput target: ~10 million tonnes per year
• Track length: 128 m main acceleration (256 m including deceleration section)
• Launch speed: 1.6 km/s (just below lunar escape, ~2.38 km/s; trajectory uses lunar gravity assist + downstream maneuver)
• Acceleration tolerance: ~1000 g for monolithic rock projectiles
• Kinetic power: 450 MW (assuming 90% driver efficiency)
• Peak instantaneous power: 16 GW at track midpoint during launch
• Average power consumption: 1.75 MW per meter of track
• Reactor cost: $2–4B on Earth; lunar-built ~10× Earth ops cost
• Power rate assumed: $2.50/kWh (10× typical US 2026 grid rate)
• Annual revenue (assumed): $100B/year per mass driver

assumed-product-price

Direct quote: "$10/kg is the assumed price for 'rocks in lunar orbit' delivered to customers." This is the product price at which the throughput economics close, not a derived breakeven. Handmer frames it as the price below which lunar bulk material isn't a commodity worth bothering with for satellite/space-economy customers.

relevance-to-q2

The Handmer piece is the cleanest first-principles lunar-mass-driver design exercise in the recent literature. It does NOT directly model lunar-surface-to-LEO cost-per-kg. The output state is "rocks in lunar orbit" (specifically, projectiles on a trajectory that catches at some downstream collection point, typically near EML2 or in low lunar orbit). Lunar-to-LEO conversion requires an additional propellant-using maneuver and a separate cost stack.

Key quote on commercial limits: "At even $500/kg, launch cost is only 5% of the total satellite deployment cost, so a lunar mass driver is unlikely to drastically improve the economics of space-based AI."

anchors-for-our-calc

• 200 kg per shot, 1 t/3 s → ~30,000 t/year of throughput per single track at full duty cycle. Handmer pushes to 10M t/yr via multiple parallel tracks or higher cadence — large scale.
• 450 MW kinetic power → ~12.5 MJ/kg kinetic per shot (at 1.6 km/s, KE = 0.5·v² = 1.28 MJ/kg, so ~2.4 MJ/kg with driver losses matches the conventional Lunarpedia figure).
• Capital structure: nuclear reactor + track. Per-kg cost dominated by capital amortization and lunar-construction premium, not energy.

limitations

• Single-blog source, not peer-reviewed
• Product price $10/kg is assumed, not derived from a cost stack
• The 10× Earth-ops-cost lunar premium is a thumb-in-the-air estimate
• Doesn't address LEO delivery — only "rocks in lunar orbit"