How the Starlink Business Became SpaceX’s Revenue Engine

For most of its history SpaceX was known as a rocket company, and its finances were private. The IPO prospectus it filed on 20 May 2026 confirmed what analysts had long assumed: the Starlink business, not launch, pays the bills. A satellite broadband service that opened to the public only in late 2020 now produces around 60% of group revenue and nearly all of its segment profit. This article explains how Starlink was built and financed, how it makes money, what the filings disclosed, and where the business stood as of October 2026. It is part of the Kurums Space Economy hub, our guide to the companies and revenue models of the space sector.

How did Starlink start and how was it financed?

SpaceX announced its satellite internet plans in January 2015. It filed for regulatory approval in 2016, launched two prototypes in 2018 and began deploying production satellites in May 2019. The build-out was financed inside SpaceX, through the parent company’s private funding rounds and launch income.

The regulatory groundwork came first. On 15 November 2016 SpaceX applied to the US Federal Communications Commission for a system of 4,425 satellites. The FCC authorized it in March 2018, and separately approved 7,518 additional satellites in a different frequency band that November. Two test satellites, Tintin A and Tintin B, launched on 22 February 2018.

Deployment began on 23 May 2019, when a single Falcon 9 carried 60 early-version satellites. From then on Starlink had a structural advantage: its owner was also its launch provider, so satellites flew on SpaceX’s own rockets instead of on launches bought from a third party.

Because Starlink was never a separate company, there is no Starlink funding history in the usual sense. Its capital came from SpaceX. That changed in form, though not in substance, when SpaceX itself went public in June 2026.

How does the Starlink business make money?

Starlink makes money mainly from monthly subscriptions paid by households and businesses for broadband delivered through a user terminal. It adds revenue from terminal sales, mobility services for aircraft and ships, a direct-to-cell mobile service sold with phone operators, and government contracts.

The consumer product launched as a public beta in October 2020 at $99 a month plus $499 for the dish. By June 2021 it had 69,000 active users in 11 countries. The appeal was simple: usable broadband in places where cable and fiber did not reach. Rural homes, remote businesses, boats and recreational vehicles were the early market.

Higher-value segments followed. US regulators authorized terminals on moving vehicles, vessels and aircraft on 30 June 2022, opening maritime and aviation. With its second-quarter 2026 results SpaceX said it had signed a major deal with American Airlines and activated service on Southwest, Virgin Atlantic, Iberia and Aer Lingus, while estimating that its penetration of the aviation market was still under 10%.

Direct-to-cell is the newest line. SpaceX and T-Mobile announced a partnership on 25 August 2022, the first six direct-to-cell satellites launched on 2 January 2024, and T-Mobile’s commercial service started in July 2025. The IPO filing reported 7.4 million monthly unique devices using Starlink Mobile across about 30 countries. Here the customer is usually the mobile operator, which pays for coverage in areas its towers do not reach.

Government is the fourth pillar. SpaceX reported more than $6 billion of contracts for Starshield, the government version of the network. Broadband Breakfast noted that government customers accounted for about 20% of SpaceX’s total 2025 revenue across all segments.

How fast have Starlink subscribers and revenue grown?

Subscribers roughly doubled each year: 2.3 million at the end of 2023, 4.4 million at the end of 2024 and 8.9 million at the end of 2025. Connectivity revenue reached $11.4 billion in 2025, up about 50% on the year before.

Growth continued into 2026. Subscribers reached 10.3 million across 164 countries and territories at the end of March, and 12 million by early June. SpaceX added 1.7 million subscribers in the second quarter alone. The IPO materials projected growth from about 10 million to nearly 17 million during 2026.

Revenue has grown more slowly than the subscriber count because the average subscriber pays less than before. Average monthly revenue per user was $99 in 2023 and $66 by March 2026, where it stayed in the second quarter. The decline has come as the service expanded to 164 countries and territories, many with lower price points than the United States. The logic of the trade is straightforward: the satellites pass over every country regardless, so capacity over a low-income region is worth filling at a lower price.

In the second quarter of 2026 the Connectivity segment reported revenue of $4.3 billion, up 66% year on year, and adjusted EBITDA of $2.6 billion. As a rough check, 12 million subscribers at $66 a month comes to about $2.4 billion a quarter. The gap to $4.3 billion suggests that a large part of segment revenue comes from sources other than standard subscriptions, such as hardware, enterprise, mobility, mobile and government services. That is our arithmetic, not a company disclosure, and the sources we reviewed did not give the breakdown.

What did the SpaceX IPO filing reveal about Starlink’s profitability?

The filing showed that Starlink is the only clearly profitable segment. In 2025 Connectivity produced $11.4 billion of revenue and a segment profit of about $4.4 billion, while the Space segment lost $657 million and the AI segment lost several billion dollars.

Group adjusted EBITDA was $6.6 billion in 2025, and long-term debt stood at $29.1 billion in March 2026. The AI segment exists because SpaceX merged with xAI in February 2026. Its losses and its capital needs are the main reason a company with a highly profitable connectivity arm reported a group net loss.

The rocket business looks different in this light. The Space segment’s revenue counts launches and services sold to outside customers such as NASA. Its largest job, launching Starlink satellites, is an internal cost. In effect the launch business has become infrastructure for the connectivity business, which is the reverse of how most people thought about SpaceX a decade ago.

The pattern held in the first report as a public company. For the second quarter of 2026 SpaceX reported total revenue of $7.8 billion, up 92%, and a net loss of $541 million. Connectivity contributed $4.3 billion, the Space segment $962 million with negative adjusted EBITDA of $205 million, and AI $2.6 billion. Readers interested in a pure launch and spacecraft company can compare these figures with our Rocket Lab company story.

How does the IPO change the picture for Starlink?

The IPO gave SpaceX a very large pool of capital and, for the first time, a duty to report Starlink’s numbers every quarter. SpaceX priced its offering at $135 a share on 11 June 2026 and began trading on Nasdaq under the ticker SPCX on 12 June.

The offering valued the company at about $1.77 trillion. SpaceX said in its second-quarter report that it sold about 639 million Class A shares for net proceeds of approximately $85.7 billion, with the offering closing on 15 June. The shares closed their first day at $161. They later fell sharply, and by late July had lost roughly half their value from the peak.

Starlink was not listed separately. Investors who want exposure to Starlink must buy all of SpaceX, including the launch business and the AI segment. The prospectus stated that SpaceX would remain a controlled company and did not expect to pay dividends in the foreseeable future.

The use of capital matters for anyone assessing Starlink. In the second quarter of 2026 SpaceX’s capital expenditure was $18.4 billion, of which $15.8 billion went to AI infrastructure. Starlink’s profits and the IPO proceeds are funding a build-out in a different industry, and the share price reaction after the first earnings report suggested that some investors were uneasy about that allocation.

What is Starlink spending on next?

Starlink is spending on three things: a new generation of much larger satellites, radio spectrum for its mobile service, and the Starship rocket needed to launch both at scale. Each is a multi-billion dollar commitment.

The new satellites are called V3. Each is designed for about 1 terabit per second of downlink capacity, roughly ten times the current generation according to SpaceX. On 28 September 2026 the fourteenth Starship test flight reached orbit and deployed 26 operational V3 satellites, the first time Starship had delivered payloads to orbit. The existing constellation numbered about 9,600 satellites at the time of the IPO filing. In January 2026 the FCC authorized 7,500 additional second-generation satellites.

Spectrum is the second investment. In September 2025 SpaceX agreed to buy EchoStar’s AWS-4 and H-block licenses for about $17 billion, split between $8.5 billion in cash and $8.5 billion in SpaceX stock, and to fund about $2 billion of interest payments on EchoStar debt through November 2027. The IPO filing put the expected closing in November 2027. The deal also lets EchoStar’s Boost Mobile subscribers use the next-generation direct-to-cell service, and SpaceX has said the next generation of satellites will deliver far more throughput than the current one. SpaceX had more than 600 direct-to-cell satellites in orbit at the time of the deal.

These projects explain why a profitable segment still sits inside a company with $29.1 billion of long-term debt. Satellites in low orbit have limited lifetimes and each generation is eventually replaced, so capital spending does not stop. The return depends on each new generation adding much more sellable capacity per dollar than the last.

What are the main risks to the Starlink business?

The main risks are falling revenue per user, permanent capital intensity, dependence on Starship, regulation in each national market, and the fact that Starlink’s cash is controlled by a parent with other priorities.

The first two are linked. If average revenue per user continues to fall while replacement and expansion costs stay high, margins will narrow. Starlink’s answer is higher-capacity satellites and higher-value customers such as airlines, shipping, governments and mobile operators. The aviation figure SpaceX quoted, penetration under 10%, shows there is room, but these customers negotiate hard and have alternatives.

Starship is a technical dependency. SpaceX is deploying the V3 satellites on Starship, so the capacity plan relies on that rocket flying often. The September 2026 flight was a milestone, but one engine failed during ascent and the mission was shortened.

Competition is developing. A Telesat executive quoted by TelecomTV described the low Earth orbit market as very competitive, and other constellations are being deployed. There is also a longer-term environmental and regulatory cost of operating thousands of satellites, a subject covered from the other side in our Astroscale story on the business of space debris.

What can founders and operators learn from Starlink?

The lesson is vertical integration used to create a cost position that competitors cannot buy. SpaceX built reusable rockets first, then used the spare launch capacity to deploy a product with recurring revenue that no rocket customer could have afforded to build.

A second lesson is about pricing for a fixed-cost network. Starlink accepted a lower average revenue per user as it expanded, because unused capacity over any region earns nothing. Subscribers grew from 4.4 million to 12 million in about 18 months while that average fell.

The third is about financing. Starlink never raised money in its own name. It grew inside a parent that could fund it through lean years, and it now funds the parent’s next ventures in return. Whether that is good for Starlink’s own customers and investors is the open question of the SpaceX IPO. Other models of selling data from orbit are described in our Planet Labs company story, and recent capital raises across the sector are listed in our space funding rounds archive.

Frequently Asked Questions

Sources

• Via Satellite: SpaceX’s IPO filing gives first look into company’s financials
• Broadband Breakfast: SpaceX reveals that Starlink is paying the bills
• SpaceQ: SpaceX reports first earnings since going public
• The Register: SpaceX and EchoStar $17 billion spectrum deal
• TelecomTV: SpaceX deploys V3 Starlink satellites as Starship reaches orbit
• Wikipedia: Initial public offering of SpaceX
• FindCheapBroadband: Starlink history timeline 2015 to 2026