Induced Demand From Cost Collapse

The market is not a fixed quantity

The standard error is treating the size of a market as a fixed property of that market. Eric Jorgenson, compiler of a book built from Elon Musk's own words, frames the mistake directly: "people have this perception that the size of a market is always going to be the size of the market." Elon Musk's counter is that a market is a function of price, and at two orders of magnitude cheaper, a large set of previously invalid use cases becomes valid. This is induced demand in the traffic-engineering sense, applied to capability rather than roads: supply at a new price point creates the demand rather than merely serving demand that already existed.¹

Volume production forces you to invent demand

The sharper half of the claim inverts the usual causation. Reaching volume production of rockets was a core part of SpaceX's strategy, not a consequence of it: the company needed iteration and volume both to cover fixed costs and to run its failure loop fast enough for the design to keep improving. Rather than launch cadence rising to meet existing demand, SpaceX had to invent new use cases to justify the volume it needed anyway, which means the question of what to do with thousands of new tons of launch capacity a year had to be answered by the company itself. The first answer was Starlink. The next ones under discussion are orbital solar compute and further space infrastructure. Jorgenson's frame for the sequence is stacked S-curves, each one a market that could not have been served at the previous cost point.¹

The frontier is a vantage point

The most portable claim is positional. Being on the frontier is not only about staying ahead of competitors; it is what makes the next opportunity visible at all. Starlink was not a bad idea in 2002, it was not an available idea: a constellation of roughly 9,000 satellites is not conceivable at two to four launches a year, and becomes straightforward at a launch every two days.¹ Nobody could have seen that opportunity from where SpaceX started, no matter how clever they were, because the thing that reveals it is a cost curve that had not yet been built.

That is why Musk did not stop at the Falcon 9 rocket. He could have sat on that position and, in Jorgenson's words, raked in cash for decades with no credible competitor. Instead he was already developing Starship before the Falcon 9 position was obviously secure. Under this model that is not restlessness, it is the only way to keep the vantage point that reveals the next business.

Open question

Induced demand is presented here as close to a law, with cases supplied only where the demand actually showed up. The failure mode this leaves open is a company that collapses its cost, builds capacity on the assumption that use cases will materialize, and finds the elasticity was lower than assumed. Even the Starlink example cuts in that direction as much as it supports the thesis: at one internal point Starlink was reportedly ten times too expensive and being built at a tenth of the required rate,¹ meaning the induced demand was real but the cost collapse that made it viable did not arrive on its own. It had to be forced a second time, by the same methods that produced it the first time.