The McCalip Calculator — Independent Reference
The only public economic analysis with interactive sliders, referenced by The Economist, IEEE Spectrum and TechCrunch. Not without limits — but honest about its assumptions.
| Scenario | Orbital ($Bn) | Ground ($Bn) | Ratio | Required conditions |
|---|---|---|---|---|
| Baseline actuel (Falcon 9, $3,400/kg, 37 W/kg) | 51,1 | 15,9 | 3,2:1 | None — available technology |
| Starcloud optimistic (70 W/kg, $5/W sat, $500/kg launch) | 16,7 | 15,9 | +5% | 3 conditions not yet met |
| Aggressive (Starship $200/kg + 70 W/kg + $5/W) | 12,1 | 15,9 | –24% | Starship operational + above assumptions |
| Aggressive + GPU 5% failure | 11,1 | 15,9 | –30% | All bets won simultaneously |
| ADS-B NETWORK Rev.4 (real operational costs included) | 20–44 | 15,9 | 1,3–2,8:1 | Economist parity disappears |
| Solar ground / outside US (possible reduction) | — | 13.9–14.9 | Moving target | India, solar zones, cheaper energy |
Collision avoidance fuel · SSA/LIDAR/FMCW sensors · ADS-B/DO-260D transponders · FCC/ITU legal costs · Launch insurance · LEOP · Full ground optical link. These items add $4–27 Bn to the optimistic scenario.
The 3 Simultaneous Conditions for Parity
The "+5%" scenario requires all three conditions to be met simultaneously. If any one fails, back to 3:1.
Falcon 9 actuel : $3,400/kg. Required reduction: ×34. Starship vol test n°12 : March 2026. Estimated commercial cost if successful: $100–200/kg. Time to commercialization: 2–4 years post-success.
Current production record: 32–34%. Required: 40%+ to reach 70 W/kg. Concentrator cells (CPV) reach 46% but mass and cost are prohibitive. Solestial (space silicon): 28–33% announced.
Starcloud-1: 11 months nominal lifespan. Actual GPU result: 5% failure/yr (better than estimated). Over 5 years = cumulative ~23% defective GPUs. 4nm GPUs remain more sensitive than military radiation-hardened chips (250nm).
The Hidden Costs — Not Modeled by Anyone
| Omitted cost item | Per satellite | Constellation 88K ($Bn) | Included in McCalip/Economist |
|---|---|---|---|
| ADS-B / STM transponders (DO-260D) | $50K–200K | 4.4–17.6 | 🔴 Absent |
| Collision avoidance fuel (extrapolated from Starlink) | $10K–50K/an | 0.9–4.4/yr | 🔴 Absent |
| SSA sensors (LIDAR + FMCW + optical) | $100K–500K | 8.8–44 | 🔴 Absent |
| Full ISL optical links (terminals + ground) | $200K–2M | 17.6–176 | ⚠️ Partiel |
| MOC 24/7 operations 88K sats (shared) | — | 0.1–0.5/yr | 🔴 Absent |
| Launch insurance + LEOP + qualification | $1,5–11M/sat | 132–968 | ⚠️ Underestimated |
| Estimated TOTAL | — | $75–640 Bn | Hors budget Starcloud |
"We always take for granted that Starship's cost is going to be hundreds of dollars per kilo. People are not taking into account the satellites are almost $1,000 a kilo right now." (TechCrunch, Feb. 2026)
Who Really Benefits?
- 🚀 SpaceX — $25M+ per datacenter launched
- 🟢 NVIDIA — space GPU validation
- 💼 Starcloud founders — $34M raised
- 🏛️ In-Q-Tel (CIA) — sovereign compute
- 📈 SpaceX IPO narrative June 2026
- 🔮 Google (Starcloud investor)
- 🔮 A16Z Andreessen Horowitz
- 🔮 Current edge/SAR clients
- 🔮 Partner satellite operators
- 📉 Retail investors (hype valuations)
- 🌍 Environment (debris + launches)
- 🇪🇺 EU clients (CLOUD Act)
- 🏢 AI users (higher costs)
- 🏛️ Taxpayers (public R&D)