Latest / The Operations Podcast with Fexingo: Process, People, and Profit in Modern Business / How SpaceX Flipped Rocket Manufacturing Operations
Transcript
- Lucas: You know the standard narrative about SpaceX — that they succeeded because of reusable rockets and a willingness to blow things up on the way to Mars. Luna: Right, the engineering story. Lucas: But there's a quieter, maybe more important story hiding underneath that. It's about manufacturing operations. How they actually built the things. And I think it's the part of the SpaceX playbook that most companies — not just rocket companies — could learn from. Luna: I'm intrigued. Walk me through it. Lucas: So let's start with a single number that tells you something is profoundly different. In the early 2000s, when SpaceX was building the first Falcon 1, they had maybe 200 employees and were producing one rocket every couple of years. By 2024, they were producing a Falcon 9 roughly every two and a half days. Luna: That's insane for a 70-meter tall rocket with nine engines. Lucas: Exactly. And by early 2026 — so this year — that cadence has accelerated further. They've now launched over 400 Falcon 9 missions total, with many of those boosters flying 20 or more times. The production system had to scale with that reuse rate. Luna: So it's not just designing a reusable rocket — it's building a factory that can churn them out like aircraft. Lucas: That's the insight. And the factory is the story. SpaceX's main manufacturing facility is in Hawthorne, California, right next to Los Angeles International Airport. It's a massive building — about 550,000 square feet. But what's remarkable isn't the size. It's the layout. Luna: How so? Lucas: Elon Musk and his team deliberately placed the factory right next to the design offices. The engineers are essentially in the same building as the production line. That's a break from aerospace tradition, where you design at one site — say, Boeing in Everett or Lockheed in Denver — and then you ship the designs to a separate manufacturing plant, often in a different state. Luna: Right, that's how the old primes work. Huge geographic separation. Lucas: That separation creates delays. A design change takes weeks to propagate to the factory floor. At Hawthorne, an engineer can walk 30 steps from their desk to the assembly line, talk to the technician, and see the problem firsthand. The iteration loop goes from weeks to hours. Luna: That's a classic lean manufacturing principle — reduce the distance between design and production. Lucas: Exactly. Toyota does this too, but in aerospace it was radical. The second manufacturing innovation is what they did with the production line itself. Traditional rocket manufacturing is what I'd call 'cathedral building.' You have one rocket in the middle of a clean room, surrounded by technicians in bunny suits, taking months to assemble it by hand. Luna: Like you're building a one-of-a-kind artifact. Lucas: SpaceX rejected that. They built a moving production line — a literal conveyor system — for Falcon 9. The rocket moves down the line, stage by stage, and workers add components at each station. It's more like an automotive assembly plant than a rocket factory. Luna: And they ditched the clean room too, right? Lucas: They did. That's another famous decision. Early on, they realized that the obsessive clean-room standards added huge cost and time with marginal benefit for their components. So they relaxed it. They still control contamination, but not to the level of legacy primes. That alone cut production time significantly. Luna: It must have felt sacrilegious to aerospace engineers. Lucas: Oh, absolutely. But the numbers backed it up. They were producing reliable rockets faster and cheaper. And that brings us to vertical integration. SpaceX makes about 80 to 90 percent of its rocket components in-house. Engines, avionics, fairings, even the seat cushions. Luna: That's the opposite of what companies like Boeing do, where they outsource huge chunks. Lucas: Right. And the standard operations wisdom is that vertical integration is risky — it ties up capital, reduces flexibility. But for a company that needs to iterate fast and control quality, it's actually a massive advantage. When a supplier can't deliver a part on time, SpaceX doesn't wait. They design and build their own version. Luna: Like the valve story. I've heard that when their supplier of a certain valve couldn't meet the schedule, SpaceX just designed and built a better one in a few weeks. Lucas: That's exactly right. And that kind of thing happened over and over. The result is that their supply chain risk is dramatically lower. They don't get held hostage by a single vendor. And because they control the whole production process, they can optimize across the entire system, not just their piece of it. Luna: So what does that mean for cost? I know the headline is that Falcon 9 costs about $67 million per launch, which is already half of what competitors charge. But what's the per-kilogram cost? Lucas: It's about $2,600 per kilogram to low Earth orbit. For comparison, the Space Shuttle was around $54,000 per kilogram in today's dollars. The Delta IV Heavy, which is a current competitor, is around $13,000 per kilogram. So SpaceX is roughly five times cheaper than its closest rival. Luna: And that's just Falcon 9. Starship is supposed to bring it down even further. Lucas: Starship is the next frontier. And the manufacturing approach there is even more aggressive. They're building Starship in Boca Chica, Texas, at a facility they call the 'Starfactory.' It's designed to produce one Starship per week eventually. Luna: One per week. That's not a rocket factory — that's a car plant. Lucas: Exactly. And they're already iterating fast. In 2025 alone, they built and tested multiple Starship prototypes, each one incorporating lessons from the previous flight. The production line itself is a prototype. They don't freeze the design and then build 50 identical units. They build, test, learn, change the line, build again. Luna: That's the opposite of the traditional aerospace approach, where you spend years designing, then build 30 identical units over a decade. Lucas: Right. And I think this is where the operations lesson for non-rocket companies really crystallizes. SpaceX treats manufacturing as a design problem — not a separate execution problem. They don't throw the design over the wall to manufacturing. The factory is part of the engineering process. Luna: That's a huge mindset shift. Most companies have a product development team and a manufacturing team that barely talk. Lucas: And the results speak for themselves. SpaceX now has a valuation that reportedly exceeds $200 billion. They launch more mass to orbit than the rest of the world combined. And they've done it by rethinking not just the rocket, but how the rocket is built. Luna: So what's the one takeaway for a listener who runs a factory, say, making furniture or electronics? Lucas: I'd say it's this: Look at the distance between your design team and your production floor — literally and organizationally. If it's more than a short walk, you're losing iteration speed. And look at your supply chain. If you're dependent on a single supplier for a critical component, consider whether you could bring that capability in-house. Not for everything, but for the parts that determine your speed and quality. Luna: And be willing to break the rules of your industry. The clean-room thing was heresy, but it worked. Lucas: Exactly. The operational playbook SpaceX wrote is not about rockets. It's about speed, iteration, and control. And that's a playbook any business can steal. Luna: Speaking of things that work — I want to mention something quickly. If you've been getting value out of these conversations, and you've found yourself applying a concept from a past episode at work, there's a small way to help keep this thing going. Lucas: Yeah, we keep it ad-free. No sponsors, no interruptions. And the way that works is a handful of listeners chip in a few dollars a month on buy me a coffee dot com slash fexingo. It's that simple — and it genuinely funds the time we spend researching and recording. Luna: No pressure at all. Just if it's made a difference for you, that link is there. And now back to rockets. Lucas: So let's zoom in on one more specific operational move: how SpaceX handles engine production. The Merlin engine that powers Falcon 9 is produced at a rate that far exceeds any other liquid rocket engine in history. Luna: How many per year? Lucas: In 2024, they were producing about 400 Merlin engines annually. That's more than one per day. For context, the RS-25 engine that powered the Space Shuttle was built at a rate of maybe two or three per year. The Merlin production line uses automated welding, robotic assembly, and a continuous flow system. Luna: And they reuse them, too, which changes the demand dynamic. Lucas: Right. Because each booster flies multiple times, the total number of new engines needed per launch is lower. But the production line still has to feed the growth of the fleet and the testing program. And they've applied the same principles to the Raptor engine for Starship, which is even more complex — it's a full-flow staged combustion engine. Luna: And that one has had its own manufacturing challenges, right? I recall Musk saying Raptor production was the main bottleneck for Starship. Lucas: Yeah, in 2021 and 2022, Raptor production was struggling. They were building maybe one per week. Then they redesigned the engine and the production process — simplified the design, reduced the number of parts, improved the supply chain — and by 2024 they were up to one Raptor per day. By 2026, they're targeting multiple per day. Luna: That's the learning curve in action. Each engine gets cheaper and faster to build as they gain experience. Lucas: Exactly. And the data confirms it. SpaceX published that the cost per Merlin engine dropped by roughly 90 percent over the first decade of production, from around $2 million per engine to about $200,000. That's a classic Wright's law learning curve — every doubling of cumulative production reduces cost by a fixed percentage. Luna: So they're not just building rockets — they're building a system that gets better at building rockets with every unit. Lucas: That's the meta lesson. The factory is a product itself. And they treat it that way. They're constantly reconfiguring the line, trying new tooling, new layouts. It's never frozen. Luna: One more thing — the human side. How do you get people to work that way? Aerospace is a conservative industry full of engineers who were trained in the old methods. Lucas: It's a real challenge. Musk has talked about how he initially struggled to find engineers willing to work on the factory floor. But they've built a culture where everyone — including the CEO — spends time on the production line. The expectation is that if you design a part, you should be able to go build it. That creates ownership and rapid feedback. Luna: So the organizational structure supports the operational philosophy. Lucas: It has to. You can't have a lean, fast factory if the organization is slow and hierarchical. SpaceX is famously flat — there are very few layers between an entry-level engineer and the top. That reduces the friction in decision-making. Luna: All right, so summing up: SpaceX succeeded because they treated manufacturing as a first-class design problem, co-located design and production, went vertical on critical components, and relentlessly iterated on both product and process. Lucas: That's the operational edge. And it's a reminder that even in a field as complex as rocketry, the principles of good operations — speed, iteration, integration — still win. Luna: I'm curious — do you think other industries, say automotive, could learn from SpaceX? Or is it too specific? Lucas: I think automotive has actually been ahead in many ways — Toyota taught the world lean manufacturing. But there's a lesson in the speed of iteration. Car companies still take four to five years to bring a new model to market. SpaceX can design and fly a new rocket variant in months. That's a pace most industries haven't matched. Luna: Maybe the next big operational revolution will come from applying that aerospace iteration speed to other sectors. Lucas: That's a great thought to end on. For now, we'll keep watching the launchpad.