Why Starlink matters
Starlink is the largest satellite constellation ever flown and the first satellite-internet system to reach mass-market, profitable scale. Barely a decade after its 2015 announcement, it counts more than 12 million active customers across 160-plus countries, a scale that makes it a genuine broadband alternative rather than a niche service for the truly unreachable.
Just as important, Starlink is the financial engine inside SpaceX. It supplied 61% of SpaceX's revenue and the large majority of the company's profit in 2025, before SpaceX had ever opened its books to public shareholders. That cash is what lets SpaceX keep funding Starship, the vehicle Elon Musk has said the entire company exists to fly, which makes Starlink arguably as important an asset to have gotten right as the reusable Falcon 9 itself.
From Redmond garage project to global network
Musk announced the satellite-internet venture in January 2015 and opened a small engineering office in Redmond, Washington. The idea moved slowly at first: the two Tintin prototype satellites did not fly until February 2018, and the first batch of 60 production satellites went up on a single Falcon 9 in May 2019. Public beta launched in October 2020, marketed candidly as 'Better Than Nothing,' and by late 2023 the business had reached cash-flow breakeven, years before it disclosed a full profit-and-loss statement to the public.
Growth then compounded quickly. Direct-to-cell satellites launched in January 2024, subscribers passed 4 million later that year, and by mid-2026, after SpaceX's IPO forced Starlink's financials into the open for the first time, the constellation had crossed 12 million customers on more than 10,800 active satellites, with revenue that had grown roughly 49% year over year.
Inside the constellation: architecture and technology
Starlink satellites orbit far lower than traditional broadband satellites, roughly 550 km versus about 35,800 km for geostationary systems, which is what gives the network its defining advantage: latency in the 20-40 millisecond range instead of the 600-plus millisecond round trip of legacy satellite internet. Delivering that at scale required flying thousands of relatively cheap, mass-produced satellites instead of a handful of expensive ones, a strategy only viable because SpaceX can launch its own hardware on reusable Falcon 9 boosters for a fraction of what any other satellite operator pays for launch.
Each satellite uses steerable phased-array antennas to serve many ground terminals at once and, since the v1.5 generation, laser inter-satellite links that pass traffic between satellites in space instead of relying entirely on ground gateways, extending coverage over oceans, poles, and countries with no local ground station. The newest V3 satellites, with roughly ten times the downlink capacity of the V2 Minis they replace, are large enough that they can only be launched on Starship, which carried Starlink V3 hardware to orbit for the first time on July 24, 2026.
From broadband to a phone carrier: the business model
What began as a single consumer broadband subscription has grown into several distinct products sold to different customers at different price points: residential and RV/roam plans from $55 to $200 a month, enterprise and maritime service with dedicated-bandwidth guarantees, in-flight WiFi now flying on 42 airlines, and direct-to-cell messaging and data sold through more than a dozen mobile carriers worldwide, led by T-Mobile's T-Satellite service, which went commercial in the US in July 2025.
The newest and most ambitious move is Starlink Mobile, a plan to sell satellite-to-phone service directly to US consumers rather than only through carrier partnerships. The FCC's May 2026 approval of SpaceX's roughly $17 billion purchase of EchoStar's wireless spectrum gave Starlink about 65 MHz of exclusive, contiguous nationwide mid-band spectrum, the building block for a genuine fourth wireless carrier. AT&T, Verizon, and T-Mobile have each publicly declined a wholesale (MVNO) deal with SpaceX, which points toward Starlink building its own consumer network rather than leasing space on an existing one.
The financials behind SpaceX's biggest business
Starlink's numbers were largely a mystery until SpaceX filed to go public. The May 2026 S-1 disclosed $11.39 billion in 2025 revenue, up 49% year over year, and $7.17 billion of segment adjusted EBITDA, making Starlink both SpaceX's largest revenue segment (61% of the total) and by far its most profitable one, funding the far costlier and still lossmaking Starship program.
Average revenue per user has actually declined over time, from roughly $99 in 2023 to about $66 in the first quarter of 2026, as the subscriber base shifted toward lower-priced international markets. SpaceX responded in mid-2026 by raising most residential and aviation prices for the first time in years, from Residential MAX (up from $120 to $130 a month) to business-jet service (up from $10,000 to $20,000 a month), a deliberate attempt to rebuild ARPU ahead of SpaceX's first earnings report as a Nasdaq-listed company (ticker SPCX) after its record-setting June 2026 IPO.
The competitive field: Amazon Leo, OneWeb, and China's constellations
For most of its life Starlink has effectively had the low-Earth-orbit broadband market to itself, and it still dominates it: more than 12 million subscribers against a field of rivals that, combined, serve a small fraction of that. Its structural edge comes largely from SpaceX's own launch-cost advantage - because SpaceX flies its own reusable Falcon 9s and, increasingly, Starship, it can put satellite mass into orbit at a cost no competitor buying launches from someone else can match.
The most credible challenger is Amazon Leo (formerly Project Kuiper), which had deployed several hundred satellites by mid-2026 and began an enterprise preview ahead of a planned consumer launch later in the year, backed by Amazon's roughly $10 billion investment. Eutelsat OneWeb has instead positioned itself around enterprise, government, and defense customers rather than competing head-on for consumers, while China's state-backed constellations (Guowang and Qianfan/Spacesail) are building networks aimed largely at markets where US satellite operators face political resistance. None of these rivals yet threatens Starlink's lead, but each narrows the multi-year head start SpaceX has built.
Risks and what to watch next
Starlink's next act depends on Starship. Scaling to the larger V3 satellites, the ones capable of gigabit-class service and the bandwidth needed to hit management's roughly 25 million subscriber target by the end of 2026, requires Starship flying often and reliably, and the vehicle is still working through the growing pains that have grounded it before. A second open question is Starlink Mobile: turning newly acquired spectrum into an actual consumer wireless carrier is a very different business from selling broadband dishes, and SpaceX has not said whether it will build its own network from scratch or find another path to market.
The growth of the constellation also draws real scrutiny. Astronomers have warned for years that tens of thousands of satellites brighten the night sky and interfere with optical and radio observations, and orbital-debris researchers have flagged low Earth orbit's most congested altitude bands, the same bands Starlink occupies, as approaching the density at which collisions can cascade into more collisions. SpaceX reports that its satellites now perform collision-avoidance maneuvers on a near-continuous basis, a sign of how crowded the environment Starlink itself helped create has become.
Sources
- Satellite Today - SpaceX IPO filing reveals Starlink financials
- Yahoo Finance - Starlink surpasses 12 million customers
- Space.com - Starship Flight 13 deploys first operational V3 Starlink satellites
- SpaceNews - FCC approves SpaceX EchoStar spectrum deal
- CNBC - Amazon Leo satellite deployment progress
- Forbes - Starlink raises residential and aviation prices