A recent video on the Cheeky Pint channel includes a deep-dive conversation with Elon Musk, Dwarkesh Patel and John Collison (released February 5, 2026).
This interview includes one of the most lucid explanations of the “Carbon Fiber to Steel” pivot Elon took with the SpaceX Starship because Dwarkesh pushes him on the manufacturing and economic implications, not just the rocket science. It contextualizes the “Steel vs. Carbon Fiber” debate as a masterclass in Elon’s 5-Step Algorithm (specifically Step 1: Make the requirements less dumb):
The “Sunk Cost” Pain
One of the most human moments in this discussion is Elon describing the sheer pain of abandoning carbon fiber.
- They had already built massive, expensive composite mandrels (molds).
- They had already ordered the raw material.
- The team was “in love” with the high-tech aesthetic of black carbon fiber.
- The Lesson: The switch to steel wasn’t just an engineering challenge; it was a psychological one. It required the leadership to say, “I don’t care that we spent millions on these molds; if they are the wrong path, we scrap them today.” This is the ultimate rejection of the Sunk Cost Fallacy.
The “Counter-Intuitive” Thermal Graph
Elon often sketches in the air during interviews, and he describes it vividly here:
- Carbon Fiber: Great at room temp, but weak at high heat (resin melts) and tricky at cryogenic cold (can micro-crack/leak).
- Steel (30X): The “miracle” is that it’s the only material that gets stronger at cryogenic temperatures (holding the fuel) while simultaneously resisting high heat (re-entry).
- The Insight: He highlights that if you look at the properties at both extremes (–165°C and +800°C), steel is actually the lighter system because you can delete the heat shield on the leeward side.
“The Machine That Builds The Machine”
The choice of steel wasn’t just about the rocket; it was about the factory.
- Carbon Fiber: Requires a clean room, autoclaves, precision placement, and slow cure times. If you make a mistake, you scrap a $2M part.
- Steel: You can weld it in a tent in a muddy field (which they literally did at Boca Chica).
- Velocity: Elon explains that steel allowed them to iterate faster. They could build a tank, blow it up, sweep up the pieces, and weld a new one in 3 days. With carbon fiber, that loop would take 3 months. Innovation per unit of time is the true metric, and steel maximized that.
Cost Per Kilogram
He reiterates the brutal economics:
- Carbon Fiber: ~$135/kg (plus ~35% scrap rate).
- Steel: ~$3-4/kg.
- When you are building a “railroad to Mars” and need to build 1,000 ships, the material cost difference is the difference between a bankruptcy and a self-sustaining city.
Elon frames the steel decision not as “finding a better material” but as identifying the bottleneck. The bottleneck wasn’t the weight of the rocket (which carbon fiber solves); the bottleneck was the cost and speed of production (which steel solves).
It is a great example of his philosophy: “The best part is no part” (deleting the heat shield) and “The best process is no process” (deleting the autoclave).
You must be logged in to post a comment.