When SpaceX began launching Starlink satellites in 2019, the business case was speculative. The premise was that satellite internet, delivered by a massive constellation in low Earth orbit rather than the traditional geostationary arc, could provide speeds and latencies competitive with ground-based broadband. It would require thousands of satellites, billions in capital expenditure, and a customer base that did not yet exist.
Five years later, the premise has been validated.
The numbers
As of early 2024, Starlink has passed four million subscribers across more than 100 countries. Revenue is estimated at roughly $6.6 billion on an annualized basis. SpaceX executives have indicated the business is cash-flow positive, though the company remains private and does not publish detailed financials.
The average revenue per user runs roughly $110-120 per month for the residential product, higher for business and maritime tiers. The maritime market, where Starlink offers a genuine performance leap over legacy VSAT services, has been a significant revenue contributor.
What makes the economics work
The core advantage is that SpaceX builds its own rockets and its own satellites. The marginal cost of launching another batch of Starlink satellites is dramatically lower for SpaceX than it would be for any competitor paying market rates for launch. When the same company that owns the launch vehicle also owns the constellation, the capital efficiency changes the entire competitive picture.
The satellites themselves are also becoming cheaper. SpaceX produces Starlink satellites at a rate that no traditional satellite manufacturer can match. It builds them in Redmond, Washington, using supply chains refined over hundreds of production runs.
The competitive moat
Competitors exist. Amazon’s Project Kuiper has Federal Communications Commission approval for over 3,000 satellites. OneWeb, now Eutelsat OneWeb, has a partial constellation in orbit. Neither has come close to Starlink’s subscriber base or its operational reliability.
The moat is not the technology alone. It is the operational lead. Starlink has been iterating its satellite design for five years. It has real-world data from millions of customer connections across every climate and terrain type. It has optimized its ground software and network architecture in ways that only come from operating at scale.
The question for the next five years is not whether Starlink survives. It is whether any competitor can close the gap before Starlink’s advantages compound further.