This is a capacity model, not SpaceX guidance. It takes the $4.29 billion of Connectivity revenue SpaceX reported for 2026 Q2 to $9.04 billion in 2027 Q4 if Starship reaches weekly cadence in January and 90% of its flights deploy Starlink V3.
The capacity grows much faster than the revenue: 4.44x against 2.11x. The difference is deliberate. More bandwidth does not create customers, terminals or paying demand at the same rate.
The calculation
The constellation snapshot separates physical satellites from usable capacity:
| Fleet | Launched | In orbit | Working | On station |
|---|---|---|---|---|
| All Starlink | 12,881 | 11,102 | 11,087 | 9,803 |
| Gen2 / V2 Mini | 8,147 | 7,890 | 7,887 | 7,252 |
| Gen1 | 4,714 | 3,212 | 3,200 | 2,551 |
| V3 | 20 | 0 | 0 | 0 |
Use the working fleet, then convert generations into one capacity unit. Gen2/V2 Mini counts as one V2-equivalent satellite. We assume Gen1 carries 20% of V2 Mini capacity. The 20 V3 satellites launched on the suborbital test contribute nothing because none reached orbit:
Opening capacity = 7,887 + (3,200 × 20%) = 8,527 V2-equivalent units
A Starship carrying 60 V3 satellites, each at 10x V2 capacity, adds 600 equivalent units. With 90% of launches going to Starlink:
Capacity added in a quarter = launches × 90% × 60 V3 × 10
Assume a five-year satellite life. One-twentieth of opening capacity, 426 equivalent units, retires each quarter:
Quarter-end capacity = prior capacity − 426 + capacity added
The launch path is three flights in 2026 Q3, four in Q4 and 13 a quarter from 2027 Q1. So the first weekly quarter is:
11,454 − 426 + (13 × 90% × 60 × 10) = 18,048 equivalent units
Revenue grows with the square root of capacity — a 0.50 elasticity, which is our assumption:
Quarterly revenue = $4.29B × (quarter-end capacity ÷ 8,527)^0.50
That gives $6.24 billion in 2027 Q1. Repeat the same calculation for three more weekly quarters and capacity reaches 37,829 equivalent units in 2027 Q4. Revenue reaches $9.04 billion.
The chart puts unlike units on one honest scale. Revenue and fleet capacity equal 100 in 2026 Q2; 13 quarterly launches — weekly cadence — also equals 100. By 2027 Q4, launches remain at 100, revenue reaches 211 and capacity reaches 444.
The assumptions and why
- Starting revenue: $4.291 billion. This is reported 2026 Q2 Connectivity revenue, including consumer and enterprise Starlink, Starshield and the existing light direct-to-cell service.
- Working fleet: 11,087 satellites. The snapshot has 11,102 in orbit and 9,803 on station. We use working because satellites still moving to their assigned planes remain part of the deployable fleet; using on-station alone would treat that already-working capacity as future growth.
- Generation mix: 7,887 Gen2 and 3,200 Gen1 working. Gen2 counts at 1.0 V2-equivalent. Gen1 counts at an assumed 0.2, producing the 8,527-unit opening capacity. This conversion, not the physical satellite count, sets the growth denominator.
- Launch ramp: three, then four, then 13 a quarter. It moves Starship from roughly monthly in late 2026 to weekly from January. If weekly cadence arrives only in March, modeled 2027 Q4 revenue falls from $9.04 billion to $8.71 billion.
- Starlink gets 90% of flights. SpaceX controls the payload and has an incentive to prove an immature rocket through repeated internal missions. The other 10% leaves room for tests and external customers.
- 60 V3 satellites per launch. That is the stated design load, not a demonstrated orbital payload. The only deployment so far was 20 on a suborbital test.
- V3 carries 10x V2 capacity. Roughly 1 Tbps against 100 Gbps is the published hardware step. It is the strongest assumption in the model because it is a spacecraft specification rather than a revenue guess.
- Five-year service life. This is the midpoint of the observed four-to-six-year range and spreads replacement evenly across twenty quarters.
- Revenue elasticity: 0.50. Doubling capacity raises revenue 41%, not 100%. A 0.25 elasticity produces $6.23 billion in 2027 Q4; 0.75 produces $13.11 billion. This is the load-bearing commercial assumption.
- Capacity earns revenue immediately. Using quarter-end capacity in the same quarter is aggressive. Terminal supply, ground infrastructure or customer activation delays would move revenue to the right without changing the fleet line.
- The published model does not change. Our SPCX model builds Starlink from subscribers, ARPU and non-consumer revenue. This supply curve becomes evidence for that model only when subscriber additions and enterprise sales show that the new capacity is used.
What to watch
- Flight 14 — does it reach orbit with a surviving payload?
- V3 booster landing burn — two consecutive failures so far.
- First ship catch, then first ship reflight — the steps needed for weekly economics.
- Pad turnaround and a second operating pad — cadence needs infrastructure, not one fast vehicle.
- Falcon flight count — if it stays near 150 while Starship remains at two to six, the handoff is not happening.
- The first operational V3 payload. The number deployed on a real orbital mission decides the capacity line.
The revenue base is SpaceX's reported 2026 Q2 Connectivity segment, captured in our quarterly notes. The working-fleet count, V3 design load, V3 and V2 capacity, satellite life and flight markers remain reported or claimed inputs requiring primary-source verification before publication. The launch ramps, 90% allocation, Gen1 capacity conversion, retirement schedule, immediate monetisation and revenue elasticities are R40 assumptions. Every capacity and revenue figure after the starting quarter is R40 arithmetic, not company guidance.
