Starlink: The Customer SpaceX Built for ItselfWide moat

Starlink produced about $11.4 billion of SpaceX's 2025 revenue, roughly 61% of the total¹. Judged as a customer relationship, that would be extreme concentration. Judged correctly, it is the reason SpaceX exists in its current form.

The launch business had a chicken-and-egg problem for fifty years. Reusable rockets only pay if they fly often; they only fly often if there is demand; and commercial and government demand together never approached the volume required. Starlink solved it by manufacturing the demand. SpaceX became its own anchor customer, flew often enough to climb the reuse learning curve, and drove costs to a level at which a constellation of thousands of satellites became affordable — which then generated more launches.

The circularity is the point, and it is also the risk. Revenue from a wholly-owned customer is not evidence that anyone else wants to buy launches at these prices, and the capital consumed building the constellation is real even when the internal transfer price is not. SpaceX reported a $4.9 billion net loss on $18.7 billion of 2025 revenue².

Watch external launch revenue separately from Starlink. If third-party launch demand grows alongside the constellation, the flywheel is genuinely self-reinforcing. If Starlink becomes an ever-larger share, SpaceX is increasingly a satellite broadband company that happens to own a rocket factory — a different business, valued differently.