There is no universal “best” space solar cell. There is only a best cell for a given mission, launch price, and lifetime.
We normalize every cell to 1 m² of active area and compute a single metric:
$/W-year = (cellCostUSD + massPerM2_kg × launchCostUSD_per_kg)
÷ (powerPerM2_W × Σ_{t=0}^{L-1} ((1 − degradation) / (1 + discount))^t)The metric is the same for every cell. What changes is the mission: launch cost, lifetime, and degradation environment. When those change, the winner changes.
The cells
All figures are for bare cells, normalized to 1 m². Costs are market estimates, not quotes.
| Cell | Technology | Efficiency | Power/m² | Mass/m² | Est. cell cost |
|---|---|---|---|---|---|
| AZUR Silicon S 32 | Silicon | 16.8% | 230 W | 1.5 kg | ~$4,100 |
| Rocket Lab ZTJ | Triple-junction GaAs | 29.5% | 403 W | 0.84 kg | ~$8,200 |
| Spectrolab XTJ Prime | Triple-junction GaAs | 30.7% | 420 W | 0.84 kg | ~$9,200 |
| CESI CTJ30 | Triple-junction GaAs | 29.5% | 403 W | 0.84 kg | ~$9,000 |
| AZUR 3G30C-Advanced | Triple-junction GaAs | 29.5% | 403 W | 0.86 kg | ~$10,000 |
| AZUR 4G32C-Advanced | Four-junction GaAs | 31.5% | 431 W | 0.86 kg | ~$12,000 |
| MicroLink IMM/ELO | IMM/ELO GaAs | 33.0% | 451 W | 0.20 kg | ~$45,000 |
The metric under different missions
8% discount rate. Degradation is technology-specific and orbit-specific.
| Mission | Winner | $/W-year | Second place | Gap |
|---|---|---|---|---|
| LEO 5-year, Falcon 9 ($3k/kg) | Rocket Lab ZTJ | 6.40 | Spectrolab XTJ Prime | 5% |
| LEO 5-year, Starship ($200/kg) | AZUR Silicon S 32 | 4.70 | Rocket Lab ZTJ | 6% |
| GEO 15-year ($10k/kg) | Rocket Lab ZTJ | 4.65 | Spectrolab XTJ Prime | 2% |
| Deep space 20-year ($20k/kg) | Spectrolab XTJ Prime | 6.18 | Rocket Lab ZTJ | 0% |
Speciation
The cells are not points on a single efficiency frontier. They are speciated:
- Triple-junction GaAs is the generalist. It wins LEO at current launch prices and wins GEO and deep space outright. Its 2–3× cell-cost premium over silicon is paid back by higher efficiency and lower degradation.
- Silicon is a niche player. It only wins when launch is very cheap and missions are short. At $3k/kg, silicon’s mass penalty erases its cost advantage. At $200/kg, the launch term collapses and silicon’s low cell cost dominates.
- MicroLink IMM/ELO never wins on $/W-year. Its cell cost is too high. Its value is specific power (>1500 W/kg at cell level), which matters for mass-constrained platforms but not for cost-per-watt-year.
- Four-junction never wins on $/W-year either. It is a performance play for missions where every watt counts, not a cost play.
Interactive exploration
Open the interactive experiment to sweep launch cost from $50/kg to $20,000/kg and watch the ranking flip. The line chart shows where silicon crosses triple-junction GaAs and where MicroLink becomes competitive on mass rather than cost.
The model is a thinking instrument, not procurement data. Cell costs are estimates; the qualitative shape — silicon vs GaAs trade-off flips with launch price — is robust, but the exact break-even launch cost is sensitive to the cost assumptions.