Regarding Skylon

I wrote this in an email to ARocket a year ago after last year’s Space Access, and I still haven’t figured out any answers to it. It is in regards to Reaction Engines’ single stage to orbit (SSTO) “Skylon” hydrogen powered spaceplane: []. The project seems to be continuing on, so it remains relevant, if only to about two dozen people in the world. Begin quoting:

Running the numbers on Skylon, I’m left with the unfortunate question of “what is worth launching 7262 metric tons of at near-term launch prices?”

Take the gross lift off weight (GLOW) and $15B development cost from the recent Skylon article on, . Pretend that the GLOW is 75% hydrogen, that hydrogen is $5/kg, and they get ops down so that the cost of a flight is down to 3 times the cost of the hydrogen alone. That comes out a bit north of $3M per flight, which is a nice price for a bit more than 10 metric tons to low earth orbit (LEO).

But the development cost still sinks it. If you compare the $15B plus the per flight cost vs. $125M per 53 metric ton Falcon Heavy (FH), you have to fly 7262 metric tons of stuff into orbit before the Skylon is cheaper. Admittedly it could be much cheaper then, but over how many years and flights can you realistically distribute the development cost? If you charge $500/lb from the start then it takes 7340 flights to break even, 75419 metric tons to orbit.

Even that ‘low’ number assumes a passel of things that are extremely unlikely, such as replacement Skylons being free (at their stated lifetime it would take four vehicles to do the FH-equiv flights), and that SpaceX wouldn’t make any improvements or figure out any reusability in 138 FH flights. In reality it’d be a lot more that 7000MT before it would break even.

7000MT seems like it should be on the same order as the mass of everything ever launched into orbit put together.

I really like reusability, and certainly wouldn’t mind having SSTO spaceplanes flying around on a regular basis. But it seems like the only way it can realistically happen is to stick someone with the development cost without their consent, whether that is wiping out the first group of stockholders or sticking it on a country’s taxpayers. Call it the Concorde model. Or the Shuttle model, for that matter.


  1. rspeed reblogged this from wikkit and added:
    this is sadly true. Reaction Engines...30-year-old idea work while
  2. wikkit posted this