SpaceX Starship: failure as an accelerator

The facts

In 2019, SpaceX began building prototypes of Starship — the largest rocket ever conceived — in South Texas using stainless steel instead of lightweight composites. The choice seemed primitive compared to aerospace tradition. But it had a precise logic: steel cost 200 times less than composites and was much faster to work with.

The strategy was radical: build prototypes in weeks (not years), test them to destruction, analyze failures, apply lessons to the next prototype. Between 2020 and 2021, several prototypes (SN8, SN9, SN10, SN11) exploded during landing tests. Each failure generated data that a year of simulations could not have produced.

For context: NASA's SLS program (based on traditional contractors including Boeing) had spent over $23 billion and taken over a decade for a single uncrewed flight in 2022. Starship, with an estimated budget of $2-5 billion, had already conducted dozens of tests.

In June 2024, Starship completed its fourth test flight successfully for both the rocket and spacecraft. In October 2024, the Super Heavy booster was caught mid-air by the "mechanical arms" of the launch tower — a feat many engineers considered impossible. In 2025, SpaceX continued iterating with increasingly complex flights, consolidating leadership in the launch vehicle market.

DAMM Analysis

Delimitation (8/10): SpaceX delimited the decision clearly: "our goal is not to build a perfect rocket on the first try — it's to build the fastest possible iteration process." This delimitation changed everything: success wasn't measured by individual launches but by learning speed. Each test had specific, measurable objectives. The explicit acceptance of failure as an expected outcome eliminated the "get everything right the first time" pressure that paralyzes traditional aerospace. The delimitation isn't perfect (10/10) because regulatory risk with the FAA wasn't always adequately anticipated.

Asymmetry (8/10): The asymmetry was favorable and well understood. Cost of a steel prototype: a few million. Value of data obtained: enormous program advancement. The cost/benefit ratio of each test was radically positive: the downside was losing a rocket worth a few million, the upside was acquiring data worth months of development. Compared to the traditional approach (billions invested in a single test), the risk per dollar invested was orders of magnitude lower. The score isn't 10/10 because the environmental and regulatory risks of some explosions weren't always optimally managed.

Room to Maneuver (8/10): The strategy itself was designed to maximize room to maneuver. Each prototype was quickly replaceable. If a technical approach didn't work, SpaceX could pivot in weeks (like the switch from composites to steel, decided in months). Parallel production of multiple prototypes meant a failure didn't block the program. Starlink funding guaranteed financial autonomy from government contract timelines. Margin was reduced only by FAA regulatory constraints.

Minimum Move (9/10): This is the most remarkable aspect. The entire Starship strategy IS the minimum move applied systematically. Instead of building the complete rocket and hoping it works (like SLS), SpaceX tested one component at a time: first the tanks, then the engines, then short flights, then high-altitude flights, then stage separation, then booster recovery. Each test was the smallest move that could validate the next hypothesis. The score is 9/10 because occasionally SpaceX combined too many objectives in a single test.