What Is Starlink’s Financial Performance?

Key Takeaways

• Starlink’s Connectivity segment posted $11.39 billion in 2025 revenue, up 49.8% over the prior year.
• Connectivity is the only profitable part of SpaceX, funding deep losses in launch and AI.
• SpaceX reached 10.3 million Starlink subscribers across 164 markets by March 31, 2026.

The First Audited View of Starlink’s Financial Performance

For the year ended December 31, 2025, the Connectivity segment that houses Starlink generated $11.39 billion in revenue, the single largest source of money inside SpaceX. For most of the company’s history, that figure and the rest of Starlink’s financial performance sat behind a wall of private-company secrecy. It became public only when SpaceX filed a Form S-1 registration statement with the U.S. Securities and Exchange Commission (SEC) on May 20, 2026, ahead of a planned initial public offering (IPO) under the proposed Nasdaq ticker SPCX.

The Connectivity numbers are the part of the filing that investors have waited years to see. In 2025, the segment reported income from operations of $4.42 billion and Segment Adjusted EBITDA of $7.17 billion. Adjusted EBITDA, short for earnings before interest, taxes, depreciation, and amortization, strips out financing and accounting charges to show how much cash a business throws off from its core operations. For the three months ended March 31, 2026, Connectivity revenue reached $3.26 billion, with income from operations of $1.19 billion and Segment Adjusted EBITDA of $2.09 billion.

What stands out is the growth rate attached to those figures. SpaceX reported that Connectivity revenue grew 49.8% year over year in 2025, income from operations climbed 120.4%, and Segment Adjusted EBITDA rose 86.2%. Profit grew much faster than sales, the mark of a business gaining operating leverage as it scales. The table below sets out the segment’s 2025 results against the prior year.

This table summarizes Connectivity’s reported 2025 financial measures and their year-over-year growth.

SpaceX credits the gains to subscriber growth, rising enterprise adoption, and steady improvement in network efficiency. The filing cautions that interim results are not necessarily a guide to a full year, and that Connectivity remains subject to spectrum rules, competition, and the capital cost of building satellites. Even with those caveats, the disclosure settles a long-running debate about whether satellite broadband could be a ly profitable business at scale rather than a money pit.

How Connectivity Became the Majority of SpaceX Revenue

SpaceX now reports its business in three parts: Space, which sells rocket launches; Connectivity, which is built around Starlink; and AI, the segment added through the acquisition of xAI in early 2026. On a consolidated basis, the company generated $18.67 billion in revenue in 2025 and $4.69 billion in the first quarter of 2026. Connectivity supplied roughly 61% of the 2025 total, which is why outside analysts increasingly describe Starlink as the financial heart of the company rather than a side project of the launch business.

The split between the three segments tells a sharp story about where money is made and where it is spent. In 2025, the Space segment produced $4.09 billion in revenue but only $653 million in Segment Adjusted EBITDA and a loss from operations of $657 million, weighed down by heavy spending on the next-generation Starship program. The AI segment generated $3.20 billion in revenue against a loss from operations of $6.36 billion. Connectivity was the lone segment delivering both growth and reported operating profit.

The table below compares the three segments on the measures SpaceX disclosed for 2025.

This comparison shows reported 2025 revenue, operating result, and Segment Adjusted EBITDA for each business.

That arrangement explains a reversal in how SpaceX is understood. Launch built the company’s reputation and remains the activity people picture first, yet it no longer drives the economics. The Space segment is a scaled but lower-margin business, with growth in government and commercial contracts offset by depreciation on hardware and continued Starship spending. Starlink, by contrast, has matured into a recurring-revenue subscription business with attractive margins. The filing makes plain that the rockets increasingly exist to deploy and refresh the satellites, and the satellites pay the bills.

The Subscriber Base Behind 10.3 Million Service Lines

Ten point three million is the headline operating number. As of March 31, 2026, SpaceX reported roughly 10.3 million Starlink subscribers across 164 countries, territories, and other markets. The company counts customers in Service Lines, meaning individual instances of broadband service tied to a user terminal rather than unique people or devices, so the subscriber figure reflects active connections rather than household headcount.

Those connections run on an unusually large piece of infrastructure. SpaceX operates approximately 9,600 Starlink broadband and mobile satellites in Low-Earth Orbit (LEO), the band of space close enough to the ground to keep signal delay low. By the company’s own measure, that fleet accounted for about 75% of all active maneuverable satellites in orbit as of the end of March 2026. Service quality has risen alongside the fleet: SpaceX reported a median residential download speed of 225 Mbps during peak hours, a figure that puts the service in competition with many ground-based broadband options.

The customer base splits into distinct groups that carry different economics. Starlink Consumer Broadband covers households and small businesses. Starlink Fixed Site serves larger enterprises. Beyond those, SpaceX sells to enterprise customers across construction, agriculture, maritime, aviation, and energy, and to government customers for public services and disaster response. Enterprise and government accounts tend to pay more per connection than residential users, which lifts the blended average even as the company adds lower-priced consumer plans in new countries.

Average revenue per user, or ARPU, is the metric that ties subscriber counts to revenue. The filing defines ARPU as service revenue divided by the average number of subscribers over a period, then by the number of months. SpaceX did not publish a single clean ARPU figure in the summary section of its prospectus, and the blended number is pulled in two directions: cheaper plans in emerging markets push it down, and enterprise, maritime, aviation, and government contracts pull it up. The mix matters because a business that grows subscribers by adding high-value enterprise lines earns very differently from one that grows mainly through discounted consumer plans.

Margins That Set Starlink Apart from Its Parent Company

Roughly 63% is the Segment Adjusted EBITDA margin Starlink’s Connectivity business reported for 2025, calculated from the disclosed $7.17 billion of Segment Adjusted EBITDA against $11.39 billion of revenue. The first quarter of 2026 looked similar, with about 64% on the same basis. Margins at that level are rare for an infrastructure business and rarer still for one that was burning cash only a few years earlier. They are the reason the segment can both grow and generate surplus capital.

Operating margin tells a more conservative version of the same story. Connectivity’s reported income from operations of $4.42 billion in 2025 works out to roughly 39% of revenue, and the first quarter of 2026 came in near 36%. The gap between the EBITDA margin and the operating margin reflects depreciation on the satellites and ground equipment, which is substantial for a business that constantly manufactures and launches new hardware. Investors who prefer operating profit to adjusted figures still see a segment earning healthy money.

The contrast with the rest of SpaceX is stark. On a consolidated basis, the company reported a loss from operations of $2.59 billion in 2025 and $1.94 billion in the first quarter of 2026, because losses in the Space and AI segments more than offset Connectivity’s profit. Several outlets reported a consolidated net loss of roughly $4.3 billion for the first quarter of 2026 once interest and other charges are included, a figure that sits below the operating line in the financial statements. The takeaway holds either way: Starlink earns, and the parent company spends those earnings elsewhere.

Margin durability is the open question. SpaceX attributes its improving profitability to falling unit costs as launch becomes cheaper and satellites become more capable, plus a richer mix of enterprise and government customers. Competition from other LEO constellations and from terrestrial fiber and 5G could compress pricing in some markets. For now, the filing shows a business whose profitability improved faster than its revenue, which is the pattern that turns a capital-hungry startup into a cash engine.

Satellite-to-Mobile and the EchoStar Spectrum Deal

January 2024 marked the start of a second Starlink network, one aimed directly at ordinary cell phones rather than fixed terminals. By March 31, 2026, SpaceX operated a dedicated satellite-to-mobile constellation of roughly 650 V1 Mobile satellites, delivering text messaging, light data, and over-the-top voice services to about 7.4 million monthly unique devices across approximately 30 countries. The service works through partnerships with around 30 mobile network operators (MNOs) on six continents, letting subscribers of partner carriers connect in areas with no ground coverage.

The financial case for this business rests on a different addressable market than fixed broadband. SpaceX estimates a total addressable market (TAM), meaning the full revenue available if it captured an entire opportunity, of roughly $740 billion for Starlink Mobile, alongside about $870 billion for fixed Starlink broadband. Those are long-range opportunity estimates rather than forecasts of near-term revenue, and the company cautions that such market sizing is inherently uncertain. The mobile figure reflects the sheer number of phones in coverage gaps worldwide rather than confirmed demand.

A spectrum transaction announced during the filing window changed the outlook for this segment. SpaceX agreed to acquire AWS-3, AWS-4, and H-Block spectrum licenses from EchoStar under a License Purchase Agreement dated September 7, 2025, and amended that November. The Federal Communications Commission (FCC) approved the deal on May 12, 2026, subject to remaining closing conditions. Owning more dedicated spectrum gives Starlink Mobile the radio capacity to move beyond text and basic voice toward broadband-grade direct-to-phone service, which is the difference between a niche safety feature and a mass-market product.

Next-generation hardware is meant to unlock that capacity. SpaceX expects to begin deploying V2 Mobile satellites, designed for broadband data and connections to internet-of-things devices, on Starship in 2027. The company labels these timelines as expectations rather than commitments, and they depend on Starship reaching operational service. Should the mobile network grow into its spectrum and satellites, it would add a second large recurring-revenue stream on top of the fixed broadband base, deepening Starlink’s financial performance well beyond its current subscriber mix.

Why Starship and V3 Satellites Drive Future Unit Economics

Sixty satellites per launch is the number that links Starlink’s future profitability to a rocket that is not yet in commercial service. SpaceX expects a single Starship flight to carry up to 60 of its next-generation V3 broadband satellites, each designed to deliver one terabit per second of downlink capacity. The company describes that as roughly a twenty-fold increase in deployed Starlink capacity compared with a Falcon 9 launch. More capacity delivered per launch, at lower cost per launch, is the lever that pushes the cost of serving each subscriber down.

The economics of satellite broadband hinge on this relationship between launch cost and network capacity. Every satellite has a limited service life and must be replaced, so a constellation is a treadmill of manufacturing and launching. SpaceX expects Starship to commence payload delivery to orbit in the second half of 2026, a milestone it presents as planned rather than achieved. If Starship reaches reliable, high-cadence flight, the company can refresh and expand the constellation far more cheaply than with Falcon 9, improving margins on the same subscriber base.

Reaching that point has been expensive. SpaceX funded $3.0 billion in research and development (R&D) for Starship in 2025 and another $930 million in the first quarter of 2026, with reporting suggesting cumulative Starship spending has passed $15 billion. Those costs land in the Space segment and depress its results today, even though the payoff would show up later in Connectivity’s unit economics. This is why the segments cannot be read in isolation: Space carries the development burden that Connectivity is designed to benefit from.

The capacity story also carries strategic weight beyond cost. V3 satellites and higher launch cadence would let Starlink add bandwidth in congested markets where demand already outstrips supply, supporting both higher subscriber counts and premium-priced enterprise service. SpaceX frames Starship as the enabler of its long-term growth across connectivity and, eventually, orbital computing. For the near term, the practical question for Starlink’s finances is narrower: how soon Starship can fly often enough to replace Falcon 9 as the workhorse that keeps the constellation fresh.

How Starlink Cash Funds the xAI and Compute Buildout

Twelve point seven billion dollars went into the AI segment as capital expenditures (capex) in 2025, meaning long-term investment in facilities and equipment, with another $7.72 billion in the first quarter of 2026 alone. Those sums dwarf what SpaceX spent on its connectivity and launch infrastructure in the same periods, and they explain why a company with a highly profitable broadband arm still reports large consolidated losses. Starlink’s cash helps make that spending possible.

The structure became more pronounced after SpaceX absorbed xAI, effective February 2, 2026, in a transaction between entities under common control. xAI had earlier acquired the social platform X in March 2025. The combined AI segment lost $6.36 billion from operations in 2025 on $3.20 billion of revenue, reflecting heavy investment in data centers, processors, and model training. SpaceX has described AI as an early-stage business it intends to fund aggressively, which only works if another part of the company produces surplus cash.

The capital flows across the three segments make the dependency visible. The table below sets out reported capital expenditures by segment.

SpaceX is not relying on Starlink alone to fund the AI push. The filing discloses Cloud Services Agreements signed in May 2026 with Anthropic, an AI company, under which the customer agreed to pay roughly $1.25 billion per month through May 2029 for access to compute capacity, with either side able to terminate on 90 days’ notice. The company also holds an option to acquire the coding-tools firm Cursor at a $60 billion implied value, with about $10 billion in fees payable if the arrangement unwinds. These deals show SpaceX trying to build outside revenue and capital for AI rather than leaning entirely on Starlink, though the broadband segment remains the steadiest internal source of cash.

For investors weighing the planned IPO, this is the structural tension at the center of the story. A buyer of SPCX shares would own a piece of a profitable, fast-growing satellite broadband business bundled with a launch business carrying Starship’s development costs and an AI business absorbing enormous capital. Reported targets ahead of the offering, summarized in coverage such as CNBC’s filing report, have pointed to a valuation near $1.75 trillion and a raise of up to $75 billion, figures that were not fixed in the prospectus itself because the price range was left blank. Those numbers remain reported expectations rather than confirmed terms.

Risks That Could Reshape Starlink’s Financial Performance

Spectrum dependence sits near the top of the risk list. Starlink cannot offer service in a market without communications licenses and frequency authorizations from the FCC in the United States and from regulators abroad, and the filing warns that obtaining and renewing those approvals is complex and slow. A denial, delay, or adverse rule change in a major market could cap subscriber growth or force changes to service. The same dependence applies with extra force to satellite-to-mobile, where spectrum is the gating resource for moving beyond basic messaging.

Execution risk around Starship is the second concern that bears directly on Starlink’s economics. The company states plainly that its growth strategy depends on developing Starship at scale and increasing launch cadence, and that anomalies, redesigns, supply-chain problems, or environmental issues could delay deployment of next-generation satellites. Because cheaper capacity per launch is the mechanism that improves Starlink’s margins over time, a slip in the Starship timeline would slow the unit-cost gains the financial case assumes. The filing frames Starship milestones as targets, not certainties.

Competition and capital intensity round out the operating risks. SpaceX competes with other LEO broadband constellations, with traditional satellite operators, and with terrestrial fiber and mobile networks, and it concedes that past outperformance does not guarantee future results. Building and constantly refreshing a constellation of thousands of satellites requires sustained heavy spending, and the company notes that it carries a substantial level of indebtedness through its credit facility and a bridge loan arranged ahead of the offering. Should cash generation or financing tighten, the pace of expansion could slow.

Governance is a different category of risk that investors have flagged. SpaceX will list with a dual-class share structure in which public Class A shares carry one vote each and Class B shares carry 10, and it expects to qualify as a controlled company because Elon Musk will hold a majority of the voting power. Coverage of the filing has reported that Musk would retain roughly 85% of votes despite a smaller equity stake, though exact figures depend on final offering terms. A controlled company can opt out of certain board-independence requirements, so a buyer of SPCX would be backing Musk’s strategy with limited ability to change it. None of these risks erases Starlink’s strong reported results, but each could alter the trajectory those results imply.

Summary

The deepest lesson of the prospectus is not that Starlink makes money, which insiders long suspected, but that SpaceX has chosen to spend that money on the two least proven parts of its empire. A standalone Starlink, judged only on its $11.39 billion of 2025 revenue and roughly 63% Segment Adjusted EBITDA margin, would look like one of the most attractive infrastructure businesses to reach public markets in years. Bundled inside SpaceX, it functions instead as the balance sheet that underwrites Starship’s development and a multibillion-dollar AI bet. Investors approaching the planned SPCX listing are therefore not buying a satellite broadband company. They are buying a profitable broadband company yoked to a launch business and an AI business that consume its surplus, governed by a founder with near-total voting control. The financial disclosures finally let outsiders separate those pieces and price them on their own merits, which may prove the filing’s most lasting contribution regardless of where the shares eventually trade.

Appendix: Useful Books Available on Amazon

• Reentry
• Liftoff
• Elon Musk
• When the Heavens Went on Sale
• The Space Barons

Appendix: Top Questions Answered in This Article

How Much Revenue Did Starlink Generate in 2025?

SpaceX’s Connectivity segment, which is built around Starlink, generated $11.39 billion in revenue for the year ended December 31, 2025. That was up 49.8% from the prior year. The segment also reported income from operations of $4.42 billion and Segment Adjusted EBITDA of $7.17 billion, making it the largest and only consistently profitable part of the company.

Is Starlink Profitable?

Yes. According to SpaceX’s IPO filing, the Connectivity segment reported income from operations of $4.42 billion in 2025 and $1.19 billion in the first quarter of 2026. Its Segment Adjusted EBITDA margin was roughly 63% in 2025. It is the only one of the company’s three segments that reported an operating profit, with Space and AI both running at a loss.

How Many Starlink Subscribers Are There?

SpaceX reported approximately 10.3 million Starlink subscribers as of March 31, 2026, spread across 164 countries, territories, and other markets. Subscribers are counted as Service Lines, meaning active broadband connections tied to user terminals rather than unique individuals. The customer base spans residential users, small businesses, large enterprises, and government accounts.

What Share of SpaceX Revenue Comes from Starlink?

The Connectivity segment supplied roughly 61% of SpaceX’s $18.67 billion in consolidated 2025 revenue. The Space segment, which sells rocket launches, contributed about $4.09 billion, and the newer AI segment added $3.20 billion. This makes Starlink the financial center of SpaceX, even though the launch business is older and better known publicly.

Why Is SpaceX Reporting Losses If Starlink Is Profitable?

SpaceX reported a consolidated loss from operations of $2.59 billion in 2025 because losses in the Space and AI segments outweighed Connectivity’s profit. The Space segment carries heavy Starship development costs, and the AI segment spent more than $12 billion in capital expenditures in 2025. Starlink’s profit is effectively redirected to fund those other businesses.

What Is the EchoStar Spectrum Deal?

SpaceX agreed to buy AWS-3, AWS-4, and H-Block spectrum licenses from EchoStar under an agreement dated September 7, 2025. The FCC approved the transaction on May 12, 2026, subject to remaining closing conditions. The added spectrum gives Starlink Mobile the radio capacity needed to expand direct-to-phone service beyond basic messaging toward broadband-grade connectivity.

How Does Starship Affect Starlink’s Finances?

Starship is designed to deploy up to 60 next-generation V3 satellites per launch, roughly twenty times the Starlink capacity of a Falcon 9 flight. Cheaper capacity per launch lowers the cost of serving each subscriber and refreshing the constellation. SpaceX expects Starship to begin delivering payloads in the second half of 2026, which it describes as a plan rather than a completed milestone.

What Is Starlink’s Total Addressable Market?

SpaceX estimates Starlink’s total addressable market at about $1.6 trillion, split between roughly $870 billion for fixed broadband and about $740 billion for mobile services. These are long-range opportunity estimates rather than revenue forecasts, and the company cautions that such figures are inherently uncertain. The estimates exclude China and Russia from the global market.

Who Controls SpaceX After the IPO?

SpaceX will use a dual-class share structure in which public Class A shares carry one vote each and Class B shares carry 10 votes each. Elon Musk is expected to hold a majority of the voting power, so SpaceX will qualify as a controlled company under Nasdaq rules. This lets it opt out of some board-independence requirements and concentrates strategic decisions with Musk.

When Is the SpaceX IPO Expected?

SpaceX filed its S-1 registration statement on May 20, 2026, and applied to list on Nasdaq under the ticker SPCX. Media reporting has pointed to a possible mid-June 2026 listing, a valuation near $1.75 trillion, and a raise of up to $75 billion. These figures are reported targets, and the prospectus left the price range blank, so final terms were not yet set.

Appendix: Glossary of Key Terms

Connectivity Segment

The part of SpaceX built around Starlink, covering consumer broadband, enterprise and government service, and satellite-to-mobile. It reported the largest revenue and the only operating profit among the company’s three segments in 2025, making it the financial core of the business.

Adjusted EBITDA

A measure of operating cash generation that stands for earnings before interest, taxes, depreciation, and amortization, with certain other items removed. It is a non-standard figure companies use to show core profitability, and SpaceX reports it separately for each segment alongside official operating results.

Average Revenue Per User (ARPU)

A metric that divides service revenue by the average number of subscribers, then by the number of months in the period. It shows how much each connection earns on average and rises or falls with the mix of cheaper consumer plans versus higher-priced enterprise and government accounts.

Service Line

SpaceX’s unit for counting Starlink customers. It refers to an individual instance of broadband service linked to a user terminal under a subscription, rather than a unique person, household, or device, so the subscriber total reflects active connections instead of headcount.

Low-Earth Orbit (LEO)

The region of space relatively close to the ground where Starlink satellites operate. Its short distance keeps signal delay low compared with higher orbits, which is what allows satellite broadband to feel responsive enough for everyday internet use and video calls.

Capital Expenditures

Spending on long-term physical assets such as satellites, ground stations, factories, and data center equipment. These outlays are large for SpaceX, especially in its AI and connectivity businesses, and they differ from day-to-day operating costs because the assets are used over many years.

Satellite-to-Mobile

A service that connects ordinary cell phones directly to satellites, filling coverage gaps where there are no ground towers. SpaceX delivers it through a dedicated Starlink constellation and partnerships with mobile carriers, currently supporting messaging, light data, and voice across roughly 30 countries.

Spectrum

The range of radio frequencies used to carry wireless signals. Access requires licenses from regulators such as the FCC, and it is limited and tightly controlled, which makes spectrum both a competitive asset and a constraint on how much capacity Starlink can offer in any given market.

Total Addressable Market (TAM)

An estimate of the full revenue a company could earn if it captured an entire opportunity. SpaceX uses large TAM figures to frame Starlink’s long-term potential, but these are opportunity estimates rather than forecasts and depend on assumptions that may not hold.

Controlled Company

A public company in which one person or group holds a majority of the voting power. SpaceX expects this status because of its dual-class shares and Elon Musk’s voting control, which permits it to opt out of certain board-independence rules that otherwise apply to listed firms.