On September 28, 2019 (eleven years after the first successful launch of the Falcon 1), Elon Musk hosted a nearly two-hour webcast update on SpaceX’s Starship development plans. While the broad outlines of the company’s plan were already well known (prior speeches, Musk tweets), the webcast provided some important new details and (fantastical?) claims that, if proven true, would fundamentally alter the ~$400 billion space industry as we know it today.
Where to start? First, an acknowledgment. The space industry has a notorious reputation for overstating capabilities while understating cost & schedule. The $500 million James Webb Space Telescope was supposed to launch in 2007 but is now targeting a 2021 launch date at a cost of ~$10 billion.
While, admittedly, “space is hard,” industry participants (both commercial and government) have strong incentives to promote overly ambitious goals. For government agencies, these “stretch goals” can be aspirational for the general public while also helping agencies to secure coveted funding. For private companies, bold visions can aide fundraising efforts while also discouraging potential competitors. SpaceX, it could be argued, is the posterchild for bold predictions. With one major caveat…the company has (eventually) delivered on nearly all its promises.
Returning to the webcast, some of the key disclosures that caught our attention included:
“I’m in love with steel”
- “The best design decision on this whole thing is 301 stainless steel”
- 301 stainless steel gets stronger at cryogenic temperatures
- Strength-to-weight ratio at cryogenic temps is equal or better than advanced composites or aluminum-lithium
- Steel has a melting temperature of 1,500 degrees Celsius vs. 300 for aluminum or carbon fiber composite
- Steel costs $2,500/ton vs. $130,000/ton for carbon fiber (98% less)
- Easy to assemble, as evidenced by the fact that Mk1 was built outdoors without a factory
- Heatshield comprised of light, but crack-resistant hexagonal ceramic tiles
- Overall, steel offers “the lightest possible reusable architecture”
- Mk1 prototype will weigh 200 tons, but with serial production should be 120 tons…maybe 110 tons
- Mk 1 lift capacity to LEO of 100 tons but clear path to 150 tons (with full reusability)
- The cost of a fully reusable system is basically the cost of the propellant. For Starship, the propellant is primarily liquid oxygen (3.5 tons of O2 for every ton of CH4)
Raptor Engine Production
- Booster designed with up to 37 engines. Designed to take multiple engines out
- Booster needs at least 31 raptors to launch
- Will need to build 100 raptor engines from today until we reach orbit (~6 months)
- Current production rate of one every 8-10 days
- Expect to reach 1 every couple of days in “a couple months”
- Targeting 1 per day by Q1 of next year
- “Building ships and boosters as fast as we can”
- Improving design and manufacturing method “exponentially”
- Mk1 and Mk2 built with plates. With Mk3 and beyond, we will simply uncoil a roll of steel and run a single seam weld down the side.
- Mk2 built in a couple months
- Mk3 in 3 months
- Mk4 in 4-5 months
- Mk4 or Mk5 to orbit in less than 6 months
- Accurate to within a few months assuming continued “exponential” learning curve.
- Current total (rest of world) Earth capacity to orbit at max launch rate (including Falcon family) is 200-300 tons. (Note – our estimate is closer to 1,000 tons)
- Super Heavy Booster should launch 20/day
- Starship should launch 3-4x per day
- Assuming 3x launches/day for 365 days/yr = >1,000 launches vs. 116 (actual) orbital launch attempts in 2018
- At the above launch rate, and assuming a 150-ton launch capacity, a single Starship could launch 150,000 tons to orbit (per year). A fleet of 10 Starships = 1.5 million tons to orbit
There’s a lot of material to unpack in the summary above. But we’re writing a research brief, not a book, so for the sake of brevity, we’ll finish with a handful of observations/conclusions.
- Getting to orbit in six months seems like a major stretch, but then again, who would have thought it possible to build a 50-meter-tall steel spaceship in 4-5 months? In an open field?
- Are we doing the math right? 400-500 Starlink satellites per launch. It could be problematic for LEO broadband competitors, at least from a cost base standpoint.
- It seems unlikely that a single starship could achieve 150,000 tons to orbit. But even if SpaceX is off by one order of magnitude, Starship would be a game-changer.
- If Starship works, the world’s most successful launch system – SpaceX’s Falcon 9 – could probably be retired. What would be the implications for all those other launchers?