Report contents
Industry landscape
Where the company competes: one global launch market, many national broadband markets
SpaceX sells two things: rides to orbit and internet access delivered from orbit. The first is a global market for medium- and heavy-lift launch, concentrated today on reusable vehicles flying to low Earth orbit (LEO, orbits a few hundred to about 2,000 kilometers up). The second is retail and wholesale broadband sold country by country under national licenses and spectrum grants. The two are linked at the root: the launch business builds and replenishes the constellation, and the constellation absorbs most of the launch capacity. The company’s 2026 S-1 is its first detailed public financial disclosure, so the public reporting record is far shorter than the operating record; the five-year lens here rests on a company that was private for most of it.
The value chain is unusually short because SpaceX owns most of it. Upstream sit propellant, specialty alloys and composites, avionics components, electronics, and access to launch ranges and test facilities. SpaceX designs and builds its own engines, airframes and satellites, and recovers and reflies boosters, which few rivals do at any cadence. Downstream sit NASA, the Department of Defense and allied defense agencies, civil space agencies, commercial satellite operators, and, for Starlink, households, enterprises, ships, aircraft and governments. Being its own largest launch customer is a structural fact worth holding onto: Contrary Research notes that growth in LEO launch demand has come largely from SpaceX’s own constellation, which means a meaningful share of launch revenue and broadband revenue rise or fall together.
Geography cuts differently in the two halves. Launch is a global market in principle, but demand is not evenly spread: US national security missions are effectively closed to Chinese launch providers, and Western commercial operators buy within a narrow set of licensed suppliers. Broadband is not a global market at all. Starlink operates in each country only after regulators grant landing rights and spectrum, and the competing substitute differs by market: fiber and cable in dense suburbs, mobile in cities, geostationary satellite in remote areas. Scale economies in launch compound globally; scale economies in broadband are local to each route and each license. That distinction drives most of what follows.
Launch is close to a one-firm market at the heavy end. Contrary Research estimated that as of July 2025 SpaceX accounted for about 84% of the mass delivered to orbit and 55% of launches by count. The named challengers are United Launch Alliance’s Vulcan Centaur, Arianespace’s Ariane 6, Blue Origin’s New Glenn, Rocket Lab’s Electron and the in-development Neutron, Japan’s H3 and India’s LVM3, with China’s state Long March family largely serving a separate demand pool. Several of these are capable vehicles; none has matched Falcon 9’s flight rate, and most are expendable or only partially reusable. Public mission-by-mission share data is not broken out by payload class, so the concentration figure above is the most concrete structural evidence available.
Barriers at the reusable LEO end are high. Capital intensity is the obvious one, but the binding constraints are engine design and manufacturing know-how, launch cadence with a trained workforce and range slots, and a flight history that insurers and satellite operators weigh. Cadence is the hardest to copy: a competitor can fly a credible rocket and still lose bids for lack of schedule certainty. The barrier is lower in small launch and in bespoke mission profiles. Contrary Research argues the market will fragment by mission type, with rivals specializing in geostationary direct insertion, dedicated rideshares and unique national-security profiles. That view is plausible; it implies SpaceX holds a fortress in standardized LEO launch and faces narrower, specialized competition elsewhere.
Broadband is more contested, though not evenly. Geostationary incumbents Viasat, Hughes (EchoStar) and SES/Intelsat sell capacity and managed service, and Amazon’s Leo is the best-funded LEO challenger, still building out its constellation. Terrestrial fiber and cable set the price ceiling wherever wires already reach. The practical market boundary is therefore where wireline economics fail: rural households, maritime and aviation connectivity, disaster response and military field use. In those segments substitutes are few, which supports pricing; in dense markets Starlink competes against fiber on cost per bit and generally loses on capacity.
What would change this picture: Starship reaching routine cadence at a cost per kilogram near the company’s own target, which the S-1 frames as a reduction of 99% or more against NASA’s historical benchmark of roughly $18,500 per kilogram.
| Program | Approx. cost per kg to LEO (USD) |
|---|---|
| NASA historical average | 18,500 |
| Falcon 9, first version | 2,700 |
| Falcon Heavy | 1,400 |
| Starship, company target | 185 or below |
Those are the company’s figures as cited in its filing; the Starship line is a stated ambition, not an achieved cost. The open question is regulatory as much as technical: how many national markets will tolerate a single dominant foreign-owned broadband supplier once the service is large enough to matter politically. That answer, not rocket performance, probably sets the ceiling on the broadband half.
Competitive advantage
Does the company have a moat? Yes in launch, not yet in returns
Moat rating: Moderate. Two tests decide this: whether the core market stays concentrated, and whether the leader earns excess returns for years. SpaceX passes the first at a level few companies ever reach and fails the second on every measure in its short public record. It listed on Nasdaq on June 12, 2026, so the filing history is brief; the figures below come from reported annual results for 2023 through 2025 and the first half of 2026.
Market position: about 84% of the mass sent to orbit
Contrary Research estimated that as of July 2025 SpaceX carried roughly 84% of the mass delivered to orbit and flew about 55% of launches by count. The gap between those two numbers is the useful part: it flies the heavy payloads. The named challengers (United Launch Alliance's Vulcan Centaur, Arianespace's Ariane 6, Blue Origin's New Glenn, Rocket Lab's Electron and the in-development Neutron, Japan's H3, India's LVM3) are credible vehicles, and China's Long March family serves a largely separate demand pool. None flies at Falcon's rate, and most are expendable or only partly reusable. SpaceX ranks first by a wide margin.
The trend has run one way. Falcon 9 flew 96 launches in 2023 and 134 in 2024, roughly half of global orbital launch attempts in each year, up from a far smaller base in 2020. Public mission-by-mission data is not broken out by payload class, so no clean five-year share series exists. The movement is nonetheless all in SpaceX's favor, driven by rivals failing to reach cadence.
That matters for how the stability test should be read. A five-year share swing of more than 5 percentage points normally argues against a barrier to entry, because it usually means customers are walking. Here the swing is consolidation toward the incumbent, which is what a barrier looks like when competitors cannot deliver on schedule. Share stability: Fairly Strong, on the evidence available.
Broadband is the harder call and the weaker one. Starlink's subscriber counts and country-level revenue are not disclosed in the financial statements summarized here, and third-party estimates vary too widely to anchor a share figure. What is clear is the structure: Starlink is the leading low Earth orbit broadband operator by satellites in service and capacity, Amazon's Leo is the best-funded challenger, Viasat, Hughes and SES/Intelsat hold the geostationary and managed-service business, and fiber and cable set the price ceiling wherever wires reach. Broadband position: Moderate.
Profitability: a rising gross margin that never reaches the bottom line
| USD bn | FY2023 | FY2024 | FY2025 | H1 2026 |
|---|---|---|---|---|
| Revenue | 10.39 | 14.02 | 18.67 | 12.50 |
| Gross margin | 41.2% | 42.9% | 49.4% | 53.0% |
| Operating income | 0.51 | 0.74 | -2.59 | -2.08 |
| Net income | -4.63 | 0.79 | -4.94 | -4.82 |
| Free cash flow | 0.11 | -5.39 | -13.95 | -34.07 |
The gross line is going the right way: from 41.2% in 2023 to 49.4% in 2025 to 53.0% in the first half of 2026. That is what pricing power or cost leverage looks like before fixed costs. Everything below it consumes the gain. Operating income was negative $2.59 billion in 2025 and negative $2.08 billion in the first half of 2026; free cash flow was negative $13.95 billion and negative $34.07 billion over the same spans.
Any return measure lands below the cost of capital. In FY2024, the one profitable year, operating income of $0.74 billion against $57.1 billion of total assets works out to roughly 1.3% pre-tax, an estimate from reported figures. FY2025 EBITDA of $4.29 billion is about 4.7% of total assets, still short of any plausible hurdle rate. After-tax ROIC was negative in three of the four periods shown, and no ten-year series exists to average. The balance sheet carries the difference: $93.5 billion of cash at June 30, 2026 against $65.6 billion of total liabilities, after $100.3 billion of financing inflow in the first half.
Profitability: Weak. Funded losses are still losses, and a company earning negative returns has not yet demonstrated that its position converts into protected profit.
Overall rating: Moderate
The concentration test passes emphatically and the return test fails. Because profitability rates weak, the framework caps the overall rating at Moderate no matter how large the launch share gets. Scale in orbit is not the same thing as earning above the cost of capital. What would lift the rating is operating income turning positive while gross margin holds above 50% and capex falls as a share of revenue; the evidence so far shows the first of those three moving and the other two not yet.
Where the advantage comes from and how long it can last: reuse economics and a licensed network
Three sources do the work here, in order of strength: the cost advantage in launch, the scale economics of the constellation, and the certification-and-coverage captivity that keeps customers from leaving. The brand and the founder's public profile are not on the list.
Cost advantage in launch: fairly strong, and tied to this decade
Reuse drives it. A Falcon 9 first stage is recovered and flown again, so the original manufacturing cost spreads across more missions instead of being written off on one. Vertical integration compounds the effect: SpaceX builds its own engines, structures, avionics and satellites, so no supplier margin sits inside the price. Marginal cost per flight falls toward propellant, refurbishment, range operations and a share of fixed overhead. The filing's own cost-per-kilogram figures show the result, from NASA's roughly $18,500 historical benchmark to about $2,700 for the first Falcon 9 and $1,400 for Falcon Heavy.
That matters in this sub-industry because launch is sold on three things: price, schedule certainty and flight history. A cost lead lets SpaceX bid below rivals and still earn a contribution margin, then plow the surplus into more flights, which spreads fixed costs again and steepens the learning curve. A competitor flying an expendable or partly reusable vehicle has to recover most of the hardware on every mission. It can match the price or protect its margin, not both.
The lead should hold on standardized low Earth orbit missions through the late 2020s. Vulcan, Ariane 6 and H3 are not reusable in the way that matters for cost. New Glenn and Neutron are years from Falcon's cadence even if they fly well. The exposure sits at the top end. Starship's target of $185 per kilogram or below is a stated ambition, not an achieved cost, and the heavy-lift transition is exactly where the next-generation constellation needs the cost curve to keep falling. If Starship slips, the advantage narrows at the moment it is most needed. Fairly Strong. A production cost lead is the most copyable kind of moat, and this one is being spent as fast as it is earned.
Scale: the constellation is a fixed cost spread over subscribers
A broadband constellation is a large fixed asset on a roughly five-year replacement cycle. Each additional subscriber adds a dish, a marginal slice of bandwidth and billing cost, and little else. The gross margin moving from 41% to 53% over three years is the visible sign of that spreading.
The structural point is sharper than the arithmetic. SpaceX launches its own satellites at internal cost; Amazon has to buy launches for Leo, and has signed Falcon 9 contracts to do it. A challenger paying its largest competitor for access to orbit starts with a cost per subscriber the incumbent does not carry. That is a real advantage and it is not easily replicated, because replicating it means building both a rocket company and a satellite operator.
The limit is geographic. Broadband scale is local: a subscriber in Chile is served by capacity over Chile under a Chilean license, and scale in the United States does nothing for the cost of serving Brazil. Landing rights, spectrum coordination and ground stations are fixed costs incurred country by country. Launch economies compound globally; broadband economies get rebuilt in each market. Moderate. It lowers unit costs within a market without closing the market to a determined entrant.
Customer captivity: certification in launch, coverage in broadband
Government and civil launch customers are genuinely captive. NASA and the Defense Department qualify a vehicle over years of mission assurance, integration and insurance review, and switching means requalification and schedule risk on a mission that cannot slip. That is durable captivity, but it covers a narrow, price-insensitive slice of demand, not the commercial bulk.
Residential broadband is the opposite. A household can cancel, return the dish and sign with a rival in an afternoon, so there is almost no switching cost. What holds the customer is the absence of an alternative at the address. In rural markets, on ships and on aircraft that absence is real and durable, and it supports pricing. Where fiber passes the house, the service competes on capacity and price per bit and generally loses. National licensing limits how many entrants can appear, which cuts both ways: it protects the position and makes a dominant foreign-owned supplier a political question once it gets large. Moderate.
Network effects are weak. Satellites do not become more useful to one subscriber because another household signs up; the benefit of a bigger base is a thinner cost structure, not a better product. That is scale, not a network.
What could erode it
The likeliest erosion does not come from a better rocket. It comes from cheaper terrestrial capacity in the places where most broadband revenue sits: fiber in suburbs, fixed wireless in cities, and direct-to-cell service from mobile carriers that turns the phone in a pocket into a substitute for a dish. Amazon's Leo is the best-funded orbital challenger, and Chinese launch capacity absorbs demand that Western suppliers cannot serve for policy reasons.
Regulation is the second channel. Landing rights and spectrum are granted nationally and can be conditioned, capped or slowed. A supplier with 84% of mass to orbit and a growing share of a country's connectivity has visible leverage, and governments respond to visible leverage.
Two things that look like moats are not. Capital access, including the $100.3 billion financing inflow in the first half of 2026, buys time but does not stop a well-funded rival; Amazon has capital too. A launch record is an achievement, not a barrier, once a competitor flies often enough to be insurable.
How long it lasts
The launch advantage is durable enough to protect pricing for most of this decade, and the evidence for it is the cadence gap plus the cost per kilogram, not the market share number alone. The broadband advantage rests on coverage gaps and licenses, both of which are real and both of which are narrower than they look: the segments without substitutes are a minority of the addressable revenue, and they are the segments rivals target last.
Because gross margin rises while operating losses persist, the record is consistent with a real moat being spent on growth. It is not yet consistent with a moat that has been proven to produce returns. What would settle it: operating income turning positive with capex below operating cash flow, and gross margin holding above 50% through that transition. What would weaken it: gross margin stalling below 45% while capex outruns operating cash flow for several more years, or Starship arriving too late for the constellation that depends on it.
Generated by DeepVal AI. Data and models may contain errors or be outdated. For research only; not personalized investment advice.
Continue reading the full report
- Comparison and summaryHow Space Exploration Technologies Corp. Class A Common Stock compares with its closest rivals, and our moat rating
Free to sign up · 20 research points · unlocking this report uses 10
20 points cover 2 full reports. No card needed.
Research reflects its generation date. Preview the first two modules.
More on Space Exploration Technologies Corp. Class A Common Stock
Space Exploration Technologies Corp. Class A Common Stock
Business, moat and key questions
Open company page