Assuming 2M subscribers end of 2022, at $80/mo for ~100/100 in rural conditions in the US. That translates to:
$1.92Bn/yr in revenue
By end of 2023, they've generated $3.84Bn in revenue or 132% of the HLS downselect contract for Lunarship.
Starlink satellites have an expected shelf life of 5 years before they deorbit. Assuming 4.5 years with a 6 month exclusion for safety, with $1.92Bn annual revenue generation, they will for the life of a their current shell, generate the company: $8.64Bn.
Assuming we need to set aside $2Bn of that for shell replacement and isp overhead for management and servicing, that's $6.64Bn on the table. Add in $2.9Bn from NASA and you've got: $9.54Bn in Starship development funding.
Elon & Gwynne have said in the past that Raptor + Starship will cost between 5-10Bn to complete. They've also said in the past that the entire 12k Starlink primary shell layers will take ~10Bn to complete.
Because Starlink is recurring revenue, and F9, FH, also generate revenue, SpaceX can likely pay for full Starship and Starlink deployment through recurring revenue of these platforms by the end of the decade.
Old space is so fucked, they have no idea what's coming. Once Starlink and Starship are completed, all revenue thereafter is next-generation platform development cash.
9m ship not enough? Fund 12m. Earth based Starlink not good enough? Need to setup base stations between Earth and Moon to have SolarLinks so that Earth and Moon Starlinks can become "regions" across space? Fund it.
Ariannespace is planning on STARTING DEVELOPMENT OF A F9 clone in 2030. SpaceX likely will be self-funding a 12, 15, or 18m Starship REPLACEMENT in 2030 for 2030-2040 usage.