The "trick" here is that he is likely speaking of operational cashflows (i.e. the 'correct' metric for low CAPEX companies).
The rub, of course, is that most of the cost for something like Starlink is designing, building, and launching the satellites, which shows up in investing cashflows.
Put differently, if the company was actually generating more cash than it consumes on all levels, start the IPO right now and SpaceX will be a $1tn+ company (instead of a $100bn+ company that people expect will one day generate cash and then be a $1tn+ company).
SpaceX will never have an IPO because being publicly traded means ceding a certain degree of control. The founders' goal is to make life multiplanetary, and the typical goal of rando shareholders is to pump up the stock price. They don't want to have to deal with the latter.
They may at some point spin off Starlink and IPO that, but not until it's fully deployed and not heavily dependent on frequent, super cheap, low bureaucratic friction launch services from SpaceX.
They already have shareholders? And as you point out, if the founder really wanted to, he could have self-funded most of SpaceX (instead of buying X for example).
He's a public company CEO already - he has no qualms about taking SpaceX public.
The rub, of course, is that most of the cost for something like Starlink is designing, building, and launching the satellites, which shows up in investing cashflows.
Put differently, if the company was actually generating more cash than it consumes on all levels, start the IPO right now and SpaceX will be a $1tn+ company (instead of a $100bn+ company that people expect will one day generate cash and then be a $1tn+ company).