Megan Skiendel: David — do you trust a success claim when the thing that's supposed to be reusable is at the bottom of the Gulf of Mexico?
David Sterling: That is a very specific question.
Megan Skiendel: Flight 13. Super Heavy Booster 20 — gone, descending too rapidly, no controlled landing. And this is SpaceX launching from Starbase, Texas, posting a success banner. I need someone to explain that framing to me.
David Sterling: Frankly, I think they're not wrong about the upper stage. Ship 40 floating intact in the Indian Ocean, 20 Starlink V3 satellites deployed — first operational V3 deployment from Starship, full stop. That's real.
Megan Skiendel: Sure — but the booster is the whole bet.
David Sterling: Right. And here's the point that keeps nagging at me — Booster 20 failed, Booster 19 failed, two months apart, both into the Gulf. The FAA investigated Flight 12, identified the root causes, mandated fixes. SpaceX said they implemented them. Flight 13 launched. Same failure mode.
Megan Skiendel: Which means either the fix wasn't real, or the Raptor ignition problem is deeper than what the FAA's corrective list captured. And I don't think SpaceX wants to say that out loud.
David Sterling: The question is what the pattern predicts — and whether the cost-per-kilogram model survives another two of these.
Megan Skiendel: But that's what makes the upper-stage story so hard to read cleanly — Ship 40 did real work. Like, genuinely. Raptor relight in space, which Flight 12 couldn't do. Floating intact in the Indian Ocean — first time that's ever happened. And six of those 20 V3 satellites were carrying cameras specifically to photograph the heat shield tiles. That's engineering data, not just bandwidth delivery.
David Sterling: Real. But narrow. Think of it like a delivery truck that made it to the destination, dropped off a third of its packages, and the engine on the way home blew again. The drop-off happened. That's not nothing. But the truck was built to carry sixty packages.
Megan Skiendel: Sixty V3 satellites is Starship's design capacity.
David Sterling: Sixty. Flight 13 flew twenty, on a suborbital path. One-third utilization. The vehicle SpaceX spent over fifteen billion dollars building — on a full rapid-reuse premise — is flying at one-third payload. The Raptor relight and the tile camera data are real wins, I'll grant both. But they're wins on a vehicle operating well below its own design envelope.
Megan Skiendel: The tile camera data, though — I mean, that's actually the thing I'd watch. If Ship 40 floated intact, SpaceX gets heat shield readings they've never had before. That feeds directly into whatever comes next for the upper stage.
David Sterling: Agreed, and that's the signal worth keeping. The upper stage is solving. Raptor relight addresses one of Flight 12's two critical failure modes. Intact splashdown gives engineers post-reentry data at a fidelity they couldn't get from a debris field. Those are not small things.
Megan Skiendel: So the hype problem isn't that SpaceX called it a success — it's that the success story is only half the vehicle.
David Sterling: Exactly that. Ship 40 is advancing. Super Heavy is stuck at the same failure point two flights in a row. And the $15 billion thesis requires both halves to work — repeatedly, cheaply, fast. One half of the vehicle making genuine progress doesn't close that gap.
Megan Skiendel: And that's the take I want to push on, because the framing circulating right now — 'SpaceX is iterating, development bumps, move on' — that framing collapses the moment you lay the timeline flat. Booster 19, Flight 12, May 22nd. Five of 33 Raptor engines fail to reignite for the boostback burn. High-speed Gulf of Mexico impact, explosion. FAA investigates. Specifically names heat damage during hot staging and erroneous alarm settings on the Raptor engines. Requires corrective actions. SpaceX says fixed. Flight 13 launches — and Booster 20 has landing burn issues. Same burn sequence. Two consecutive flights.
David Sterling: Couldn't this still be a tuning problem, though? Two data points isn't a pattern — it's two data points.
Megan Skiendel: Two data points after targeted corrective action is different. That's the part that breaks the 'normal iteration' framing. If the FAA identified the specific root causes — heat damage, bad alarm settings — and SpaceX said they fixed those specific things, and then the same failure mode appears — that's not tuning. That's a pattern.
David Sterling: I'll grant that. The prior for 'we fixed it' is now falsified at least once.
Megan Skiendel: And Mark Spiegel is saying this publicly — not as a random critic, he has institutional standing — flagging that the booster failure directly undermines the economic model. He's not wrong. And here's what nobody wants to say: SpaceX announcing it's targeting a mechanical tower catch for Flight 14 is itself an admission. They're not saying 'we fixed the landing burn.' They're routing around it.
David Sterling: The Mechazilla pivot is interesting. That's — wait, is that the right read? Because tower catch was always part of the architecture.
Megan Skiendel: It was always the goal, sure. But the sequencing matters. You don't rush to the tower catch while the landing burn sequence is still unresolved unless the landing burn is harder than you modeled.
David Sterling: I'll accept the signal. I do want to be honest, though — we're reading external indicators here. We don't have internal SpaceX failure data, we don't have the actual engineering post-mortem on Booster 20. We're inferring from behavior, not confirmed failure modes.
Megan Skiendel: Fair. And what Flight 14 actually needs to prove — and what two consecutive losses do to the $15 billion model — that's a harder conversation we should get into.
David Sterling: The attrition math keeps coming up. If Super Heavy is lost or damaged on every V3 operational flight so far — and it is, both Flight 12 and Flight 13 — the $15 billion doesn't amortize. It writes down. The whole airline-economics argument for Starship requires both stages returning, repeatedly, cheaply. You've got zero for two on the expensive stage.
Megan Skiendel: And the market clocked it. SPCX down 1.5% after hours — $113.38 — after a flight SpaceX called a success.
David Sterling: That's the tell. Not a panic — but a down move on a headline win is directionally telling.
Megan Skiendel: And then three weeks before Flight 13, Falcon 9 booster B1067 — June 8th, 2026 — logs its 29th Starlink launch. Five-plus years of service. Reuse record. Same company. That's the comparison nobody wants sitting in the same paragraph.
David Sterling: It's brutal, frankly. Picture a freight carrier — their pilot on a red-eye lands the upper deck of the plane perfectly, twice. But the tail section collapses on touchdown each time. Meanwhile their own older regional jet quietly logs its 29th perfect landing that same week. The Raptor engines on Super Heavy are the tail section in that scenario.
Megan Skiendel: Which is the thing — B1067 isn't a competitor's jet. That's SpaceX's own benchmark. They know exactly what reliable reuse looks like. They built it.
David Sterling: So Flight 14 — I mean, it becomes the only data point that matters now. Not for the technology story. For the category of problem. Recovery snaps the narrative back to execution risk. Another Raptor ignition failure, another Gulf of Mexico impact — that shifts the conversation to design risk. Those are very different investment theses.
Megan Skiendel: And whoever greenlighted Flight 14 inside SpaceX — they know that. They are sitting with exactly that framing right now.
David Sterling: The fact is, two consecutive booster losses across the only V3 operational flights on record — that's not a development footnote. That's the core economic premise undemonstrated. Flight 14 either proves the model or it moves the goalposts. There's no third outcome.
Megan Skiendel: The one I keep sitting with — and I don't have an answer — is whether SpaceX actually knows the Raptor ignition problem is fixable on Super Heavy at the scale the reuse model requires. Like, internally. Not what they're saying publicly. Whether they believe it.
David Sterling: I mean — that's the question Flight 14 either answers or defers. If that booster lands, Spiegel's position becomes a footnote. If it doesn't, the conversation isn't about execution risk anymore. It's about whether the Raptor cluster on Super Heavy has a structural vulnerability that two consecutive losses didn't fix. And a $15 billion full-reusability thesis has to get repriced at that point. Not maybe. Has to.
Megan Skiendel: And we genuinely don't know what they know. Their iterative track record is real. B1067 didn't happen by accident.
David Sterling: No. It didn't.
Megan Skiendel: Flight 14 is the whole conversation now. Good to think through it with you.
