We didn’t. We didn’t pause to question the premise when Elon Musk claimed Starlink could one day carry 50% of global internet traffic. The number was intoxicating—a narrative of conquest that felt like a bull run for the sky. But sentiment is a shifting tide, not a solid ground. And as I sat in Riyadh, re-reading the transcript of David Friedberg’s podcast, I felt the familiar pull of a myth waiting to be debunked.
Context: The article is a deep analysis of Starlink’s long-term traffic and revenue predictions, based on a recent industry brief. The core claim: Starlink could generate $400 billion annual revenue and eventually 50% of all internet traffic. Friedberg, a potential investor, and Musk, the ultimate insider, painted a future where satellite broadband becomes the backbone of global connectivity. But the analysis I’ve parsed reveals a lattice of hidden assumptions—technical, economic, and geopolitical—that the hype machine conveniently ignores.
Core: Starlink’s product is a satellite broadband access service, not a traffic carrier. Its technical architecture, LEO constellation with about 7,000 satellites, offers global coverage and low latency. But to handle 50% of global internet traffic (projected ~396 EB/month by 2027), Starlink would need 15,000 to 40,000 satellites—a scale that strains spectrum availability, orbital debris limits, and ground station backhaul. The article’s “no obvious obstacles” claim is a sleight of hand. I’ve seen this before in crypto audits: the assumption that linear scaling works when physics says otherwise. The unit economics are equally fragile. Current ARPU of $100-120/month for 6 million users yields ~$60-100 billion revenue. To hit $400 billion, Starlink needs 30-35 million users—a 5x growth in a market where the “no alternative” niche is finite. The $1 trillion revenue target implies 400-600 million users, equivalent to 40-50% of the global telecom services market. That’s not a stretch; it’s a fantasy.
Contrarian: The hidden variable is the cost of capital maintenance. Friedberg’s $30 billion free cash flow estimate assumes a 75% FCF margin—unheard of in telecom, where margins are 10-20%. This assumption implicitly presumes the constellation is fully built and won’t need expansion. But the “carry 50% traffic” goal demands continuous satellite launches and replacements (5-7 year lifecycle). The math is self-contradictory. Furthermore, the real growth engine isn’t consumer broadband—it’s Direct-to-Device wholesale. But that model, while bypassing hardware costs, offers lower margins and cedes customer relationships to mobile operators. The high-ARPU enterprise clients (maritime, aviation, government) are a finite pool—global shipping has ~100,000 vessels, aviation ~25,000 planes. Even at maximum capture, that’s a few hundred billion, not a trillion. The narrative of “AI-driven bandwidth demand” ignores that most AI traffic flows inside data centers, not over satellites. In the ledger’s silence, the true story whispers: Starlink is a brilliant niche solution, not a global hegemon.
Takeaway: The Starlink narrative is a myth of unbounded scale, but the physical and economic constraints are real. Every bull run in crypto taught us that the gap between vision and execution is where the crash happens. The question isn’t whether Starlink can grow—it will. The question is whether the market will price in the $1 trillion dream before the reality of maintenance costs and competitive pressure (Kuiper, 5G FWA) erodes the fantasy. I’ve seen this pattern before. We didn’t learn from Raptor Protocol. We didn’t learn from Terra. Now, the sky is the next ledger.
