A payload reduction of 30-40% is huge. And even if refurb and recovery is less than 10%, you still loose the 2nd stage at every lunch (18% of the rocket by weight). You then have to include the extra development costs and additional risk of failure.
I highly doubt they hit break-even with only two launches, the math just doesn't add up.
If reusing rocket was so profitable, there is a high chance someone else would have done it before. Landing a rocket isn't totally new tech. NASA did it on the moon 50y ago.
After the massively incorrect analysis by popular YouTuber Thunderf00t, I made a video analyzing SpaceX's reusability https://www.youtube.com/watch?v=36o4UrS9OS4 It's 8 minutes but I recommend watching at 2 times speed.
The latest numbers are that the marginal cost of relaunch is $15 million, with $10 million of that the second stage. Refurbishment cost is 27 days with a $1 million refurbishment cost (not $250,000 as I mistakenly said in the video, see description). The net effect is the cost of marginal cost of reflight is approximately 25% but the payload reduction is 30%, which leads to a break even point of 3 launches. After 10 flights, the cost to orbit per kilogram is halved. SpaceX has demonstrated 8 flights of a single booster.
Note, these are all using SpaceX's own numbers (but so was the payload reduction and original reuse cost figures). It depends on recovering the $6 million fairing (which does not happen every mission). All numbers are sourced from an Aviation Week May 2020 podcast that I linked to. Yes, the cost numbers are from SpaceX but the only public refurbishment cost analogue we have is their 27 day turnaround time.
SpaceX are currently operating at a sixth of the rest of the industry and use their cost advantage to launch 60 Starlink satellites at their $15 million marginal cost.
The traditional military-industrial complex with cost-plus accounting is "subcontractors all the way down", and the incentives weren't aligned to make an industry dominating rocket until arguably the 1990s commercial satellite boom (eg, the original Iridium constellation). Until 2008, there hadn't ever been a privately funded rocket reach orbit ever. Until SpaceX. Prior rockets we designed, owned and approved by the US government under World War 2-era cost-plus accounting arrangements.
While the 2nd stage is 18 % of the rocket by weight, it only has one engine. The booster has nine. Engines are by far the most expensive part of the rocket, or used to be. (ULA says that engines make 65 per cent of the total cost of their first stages.)
> If reusing rocket was so profitable, there is a high chance someone else would have done it before.
AFAIK it is only like 10-12 years that computers and engineering software got good enough to simulate everything that happens in a rocket engine, sloshing of fuel etc. Cruder simulations were available before, but not good enough for that.
SpaceX engineers did most of the design and testing on computers. It took them only a few tries in the real world to nail the first landing.
Earlier pioneers would have to try in practice, which would mean a lot more failures and a lot more money, with a possible deadly incident in between.
I highly doubt they hit break-even with only two launches, the math just doesn't add up.
If reusing rocket was so profitable, there is a high chance someone else would have done it before. Landing a rocket isn't totally new tech. NASA did it on the moon 50y ago.