Firefly Space Systems charges full-speed toward low Earth orbit

How does one go about starting a space startup? A lot of YC companies are tech companies where one can go into Ruby on Rails or Java, crank something out with virtually zero capital, get the hockey stick growth and then it's a company but with space startups this definitely isn't the case. What would I need to convince investors?

Out of curiosity, which companies do you define as being in the "first", "second", and "third" rounds of space startups?

> Makes me think they are leaving something big out of the article.

I'm certainly not leaving anything out on purpose!

Speculating, though, I imagine it's a question of risk versus reward. There have definitely been advances in rocket designs since the 1960s—the RS-25 is a great example of a more modern high-efficiency, high-performance design (the MSFC propulsion engineers I interviewed a couple of years ago for my F-1 piece likened the RS-25 to a Ferrari, while the F-1 is more like a giant Mack truck).

But developing rocket engines has until relatively recently been the domain of big aerospace companies on big government contracts, and under those circumstances it's often far more short-term cost efficient to go with the sure thing and either use an existing rocket or design one with a conventional bell nozzle.

Rocketdyne in fact proposed an annular aerospike engine for the space shuttle when the program was in early development, but the technology wasn't considered mature and the RS-25 was designed instead.

R&D funding for rocketry has been very sparse, and large scale rocketry has consistently been extremely risk averse. Because of that there have only been a few different rocket engine designs which have been advanced towards production suitable for use on orbital launch vehicles. In the US, for example, there was a period of development on LOX/Kerosene engines in the US during the Apollo program but then the Hydrogen fad came along with the Shuttle program and that largely stopped until SpaceX came along. The SSME, RS-68, and SpaceX's engines are the only rocket engines with more than 500 kN of thrust that have been developed in the US in decades. That gives you a sense of how sparse the development landscape for rocket engines has been.

We simply have not explored much of the design space for rocket engines as of yet. We haven't even explored many of the most promising propellants (LOX/Methane being one example), let alone most of the rocket engine designs. Pursuing a new engine design and scaling it up to a size suitable for orbital rocketry is typically incredibly expensive and difficult, which is why it has been done so rarely.

Additionally, government backed R&D has often suffered from a bit from "Wile E Coyote" engineering. If a project works on something new and that project fails, or sometimes merely ends, then whatever was being worked on is often abandoned and treated as though it's impossible. Just look at the failure to followup on the DC-X work, despite the program overall being quite successful. Yet now there are several companies who are following up on those design principles (VTVL reusability) such as SpaceX, Blue Origin, and now Firefly. Aerospikes have a lot of potential but also a lot of difficulties, nobody has ever put in sufficient work to figure out whether those difficulties are surmountable.

Ok, on a PC so here's a little more detail:

In theory, aerospike engines are pretty efficient at nearly any altitude. It has to do with the pressure and temperature variations as the rocket travels through the atmosphere. (There are pockets of isothermal regions, or regions where the temperature is the same in the atmopshere, and there are pockets where the temperature varies; therefore the Mach number varies). The atmospheric pressure and temperature affect the flame of the aerospike and act like a "variable" nozzle.

Aerospikes aren't as popular because,

1) they're heavy. more deadweight mass on a launch vehicle is a big no no. By deadweight I mean everything on the launch vehicle that doesn't contain the a) payload b) propellant. Because once you burn out (or run out of propellant), usually that's when you stage (or separate) your payload from the launch vehicle. At that point, the launch vehicle is useless; deadweight.

2) they require a massive amount of cooling: you're basically heating up a piece of metal to launch temperatures during the launch sequence. You have to make sure your "spike" doensn't melt.

3) There is a bunch of research and development into other cheaper alternatives, and we already have cheaper and efficient alternatives. This basically pushed aerospikes to the "novelty" category. I'm sure if there was a big technological "push" for aerospikes I'm sure we could get something that's more efficient but that research and time will cost more than to just use the rockets we have now.

ADDENDUM; here's a video from Rocketdyne explaining aerospike engines: https://www.youtube.com/watch?v=EWf4iOMSPNc

You can shape Laval nozzles to have peak efficiency at a specific altitude/pressure/velocity. If you are committed to multi-stage rockets - and there are many reasons to do this - then you can have a different shaped nozzle for each stage. This can lessen the advantage of an aerospike and make the extra weight and inefficiency more of an issue.

For example, the nozzles on the shuttle's solid boosters were optimized for low atmospheric use, but they were discarded relatively quickly. The nozzles on the shuttle's main engines were most efficient out of the atmosphere.

I'm on mobile so this'll be pretty in-a-nutshell:

Aerospikes are rocket nozzles that are very efficient across a wide range of altitudes because at different altitudes you have different pressures and temperatures. Normal rocket engines usually compensate for this with variable geometry nozzles.

I think the first flight of a rocket which actually flew with an aerospike engine was in large part a university project - http://www.astronautix.com/engines/garspike.htm .

Lots of work done on aerospike engines. Google IYF. FYI, here's a video of a linear aerospike test:

https://www.youtube.com/watch?v=FcW9kUUTfxY