Starlink is SpaceX’s cash machine, but the maths is getting harder

SpaceX’s satellite internet service had 10.3 million subscribers at the end of the first quarter of 2026, more than double the 4.4 million it reported a year earlier. That growth rate is extraordinary by any standard in telecommunications, let alone for a business that delivers broadband from orbit.

But the S-1 prospectus SpaceX filed ahead of what would be the largest IPO in history reveals a less comfortable trend underneath the headline numbers. Average revenue per user fell to $66 per month in Q1 2026, down from $86 a year earlier and from $99 in 2023.

The ARPU problem

The decline is not accidental. Starlink has been expanding aggressively into price-sensitive markets across Africa, South-East Asia, and Latin America, where monthly subscriptions are set well below US rates to match local purchasing power.

The strategy has delivered volume. But it has also meant that even as the subscriber base more than doubled, operating income in Q1 barely moved, rising from $1.03 billion to $1.19 billion, according to CNBC.

For the full year 2025, Starlink’s connectivity segment generated $11.39 billion in revenue and $4.42 billion in operating profit. Those are strong numbers.

The question is whether SpaceX can keep growing the top line without the bottom line stalling.

Moving into harder territory

Starlink built its subscriber base in rural and underserved areas where it was often the only broadband option. The next phase of growth requires moving into suburbs and cities, where it will compete directly with fibre, cable, and fixed wireless providers that can match or beat it on price and reliability.

“People underestimate the ability of terrestrial competitors to respond with pricing, bundling, and the like,” telecommunications analyst Tim Farrar told CNBC. Farrar has also noted that Starlink’s user terminals cost roughly three times as much to produce as the modems used by terrestrial internet providers, limiting SpaceX’s ability to compete on equipment pricing in markets where consumers already have alternatives.

SpaceX appears to recognise the pressure. In May 2026, it raised prices on every consumer Starlink plan by $5 to $10 per month, a move that signals a shift from pure subscriber acquisition toward extracting more revenue from the installed base.

The Starship dependency

The next generation of growth depends on hardware that does not yet fly operationally. SpaceX’s Starlink V3 satellites, each capable of delivering roughly 1 terabit per second of downlink throughput compared with 80 gigabits on a V2 Mini, are too large to launch on Falcon 9.

They require Starship.

SpaceX has spent more than $15 billion developing Starship, and the vehicle completed its twelfth flight test on 22 May 2026, the first launch of the V3 variant with Raptor 3 engines. But Starship is still in testing.

SpaceX flew five Starship missions in 2025 against a target of 25.

Any delay in Starship reaching operational cadence pushes back the deployment of V3 satellites, which in turn limits the capacity expansion Starlink needs to serve tens of millions of additional subscribers without degrading service quality.

Competition is finally arriving

For most of its existence, Starlink has operated without serious competition in the consumer satellite broadband market. That is changing.

Amazon’s Leo satellite internet service, formerly Project Kuiper, entered enterprise beta in April 2026 with commercial availability targeted for mid-2026. Amazon has committed $10 billion to the programme and has signed beta partnerships with Verizon, AT&T, Vodafone, and NASA.

Amazon claims Leo will offer download speeds of up to 1 Gbps, roughly double Starlink’s typical throughput.

Amazon has also agreed to acquire Globalstar for $11.57 billion, adding terrestrial spectrum and direct-to-device capabilities to its satellite stack. The FCC granted Amazon a deadline extension on its satellite deployment requirements, giving it until 2029 to put all 3,232 planned Gen 1 satellites into orbit.

The IPO arithmetic

SpaceX priced its IPO at $135 per share on 11 June, valuing the company at approximately $1.77 trillion and raising $75 billion, the largest public offering in history. The stock is set to begin trading on Nasdaq under the ticker SPCX on 12 June.

At roughly 94 times 2025 revenue, the valuation implies that investors are pricing in years of aggressive growth across Starlink, Starship, and the xAI artificial intelligence operations that SpaceX absorbed in a February 2026 merger. Morningstar has put a fair-value estimate of $63 per share on the company, less than half the IPO price.

Starlink is the only consistently profitable division. SpaceX as a whole swung from a $791 million net profit in 2024 to a $4.94 billion net loss in 2025, driven by the costs of integrating xAI and expanding AI infrastructure.

The connectivity business, anchored by Starlink, accounted for 61% of total 2025 revenue.

The regulatory picture

SpaceX has secured significant regulatory wins. The FCC approved 7,500 additional Starlink Gen2 satellites in January 2026 and granted SpaceX a waiver to exceed legacy interference limits for its constellation within the United States.

In May, the FCC approved SpaceX’s acquisition of roughly 65 MHz of nationwide spectrum from EchoStar, giving Starlink the spectrum it needs for direct-to-phone 5G connectivity.

But the waiver on interference limits comes with a clause requiring SpaceX to cease operations immediately if harmful interference to older geostationary satellite systems is proven, a risk that incumbent operators such as SES and Viasat have signalled they intend to challenge.

The flags

The ARPU decline from $99 to $66 over roughly two years is sourced to SpaceX’s S-1 filing and corroborated by CNBC and Morningstar analysis. Tim Farrar’s claim that Starlink terminals cost roughly three times as much to produce as terrestrial modems could not be independently verified beyond his own public statements and should be treated as an analyst estimate.

Amazon Leo’s claimed 1 Gbps download speed is Amazon’s own marketing figure and has not been independently tested at scale. The Morningstar fair-value estimate of $63 per share is one firm’s assessment and does not represent a consensus view.