Source review: Musk $10/kg Starship target

Summary

Claims reviewed

Claim 1 — $10/kg target (Musk)

Quote (paraphrased from multiple Musk statements): Starship will reduce launch costs to ~$10/kg. Verdict: Different conclusion

The q1 calc cannot reproduce $10/kg from first-principles assumptions. Even the most aggressive scenario (100 reuses, 150t payload, 8% refurb, $0.5M ops) gives ~$59/kg internal cost. To get to $10/kg requires either:

• Sub-2% refurb (vs Shotwell 2017's "<50%" anchor)
• Plus near-zero ops cost
• Plus zero margin on customer-facing pricing

Or — as Codex flagged in pass-03 audit — Musk may be describing internal cost not list price. Under TAI-grade automation (compressing refurb labor 10×, cadence 10×) the internal-cost figure could approach $10-20/kg. Customer price still requires zero margin to match.

Our position: the figure is aspirational and outside the calc envelope without these extra mechanisms. Cite Musk explicitly; do not adopt $10/kg as an unattributed central estimate.

Claim 2 — "Propellant accounts for roughly one-third" of $10/kg

Verdict: Merits investigation

If $10/kg is the total and propellant is 1/3 ≈ $3.33/kg, then for 150t payload: propellant = $500k/launch. My calc uses $1M propellant per launch (Musk's own previously-stated number). The ratio implies Musk has revised either the propellant figure down or the total target up since the earlier statements.

Either way, the propellant arithmetic constrains the rest. If propellant is genuinely $500k/launch and represents 1/3 of cost, total $1.5M/launch, divided by 150t = $10/kg. This matches the target but assumes:

• Hardware fully amortised (effectively $0/launch — requires very high reuse + low refurb)
• Ops cost <$100k/launch (vs my late-era assumption of $500k)

These are aggressive but not impossible under industrial-explosion conditions.

Claim 3 — Implicit: $10/kg achievable as Starship matures

Verdict: Consistent (conditional)

Consistent with the q1 framework given: (a) sustained automation pressure on refurb + ops, (b) zero margin during market-capture, (c) reuse counts beyond 100 cycles. The published $1,600/kg early → $100-150/kg long-term projection (cited in the same source) is the more conservative reading without those assumptions.

Cross-reference

• The reconciled mapping in [pass-03-reconcile.md] places Musk's $10/kg outside the standard envelope and inside a narrow corner requiring TAI-grade compression. This source review confirms that mapping with arithmetic.
• The propellant-arithmetic claim (Musk's 1/3 ratio) is non-obvious and worth promoting to a numerical claim in claims.yaml during the write pass.
• Musk should be cited but not silently treated as ground truth (per Avi's guidance + anti-pattern #13 voice rules).

Musk's $10/kg target — attributed citation

The claim: Musk has stated $10/kg to LEO as a Starship target cost, reiterated multiple times across years. Propellant accounts for "roughly one-third" of that cost per his commentary.

Status as of 2024-2025: Aspirational. Not achieved. Starship still in test-flight phase.

Alternative public projections cited in same reporting:

"With Starship, full reusability is expected to drive launch costs down to roughly $1,600/kg to low Earth orbit, with the potential for further reduction to about $100/kg to $150/kg over time."

This second range ($1,600/kg full-reuse → $100-150/kg long-term) is a substantially more conservative projection than Musk's $10/kg headline target. It's not clear from this source who originated the $1,600 / $100-150 numbers — possibly aerospace analysts or other SpaceX personnel.

Relevance for q1 reconcile

Musk's $10/kg target is cited but should be qualified with:

• Aspirational, not achieved
• Requires further-than-Falcon-9-historical refurb compression
• Conflicts with cited "$1,600/kg full-reuse → $100-150/kg long-term" projection

The $100-150/kg long-term figure sits roughly inside my optimistic-late scenario ($59-277/kg per pass-02-calc.md). The $1,600/kg early-Starship figure is in line with my pessimistic scenario.

The $10/kg figure is outside my entire scenario envelope without additional compression mechanisms (TAI-grade automation, zero-margin pricing).