SpaceX and the categorical imperative to achieve low launch cost

SpaceX is the fastest growing space company worldwide. It has achieved major successes since its creation in 2002, and its level of technical execution has established the company as a benchmark for the newspace community and as a model to follow for countless start-ups in the sector. 

Incumbent players in the launch sector have been looking at SpaceX, first with skepticism, then incredulity, and now with awe, witnessing the company unfolding its business model, initially sparked by Elon Musk’s criticism that access to space was too expensive. 

Today it is generally recognized by all observers that SpaceX, with its partially reusable Falcon launch system, has achieved major cost reduction for orbital launch.

The same observer also notes that the customers of SpaceX, as a whole, are not yet benefiting from any significant reduction of the cost of access to space, because SpaceX is not incentivized to pass the economies they achieve to their customers, or only very marginally, if they do. 

We believe that the economies achieved by SpaceX on launch are irrevocably tied to the success of SpaceX, to the point that they provide the foundation of Starlink’s seemingly unstoppable development. 

SpaceX is a company that has four main lines of business: 

Space launch. SpaceX serves three market segments. Starlink launches (80% of the total), United States government launches (NASA and DoD) and commercial and export launches (SES, OneWeb, ESA). This launch business is usually presented as being extremely profitable because SpaceX, thanks to its reusability technology, is driving down its costs and is able to generate gross margins on every launch performed for external customers (while Starlink launches benefit from at-cost pricing with no profit made). It is generally assumed that SpaceX launch business yields about $3.5 billion in revenues in 2024 (inclusive of Dragon cargo and crew missions), with about 20 to 25 launches per year. 

Broadband service. Through its 6,000 satellite Starlink constellation, SpaceX is able to deliver global broadband access to fixed and mobile terminals around the globe. It is currently the largest provider of such service in the U.S., with a subscriber base of 3 million or more. This business is assumed to yield about $4.5 billion in 2024 (80% of this is service revenue, the rest is composed of terminal sales), though estimates vary. It was recently announced that in Q1 of 2023, Starlink achieved ”breakeven cash flow”, and that SpaceX turned a profit in 2023. Analysts usually recognize that to achieve breakeven cash flow, and turn a profit on Starlink, SpaceX must achieve a very low Capex on Starlink, so low in fact, that the only viable assumption is that each Starlink launch costs less than $50 million (satellites and launch included), which in turn leads to the assumption that Falcon 9 costs charged to Starlink must be lower than $28 million per launch, if we take at face value Elon Musk’s statements that the cost of Starlink satellites is in the order of $1,000 per kg, or about $17 to 18 million worth of Starlink satellites on each Falcon 9 launch. In fact, any assumption above the $50 million per launch would be incompatible with the statements about profit and cash flow in 2023. 

Satellite supplier. SpaceX is a supplier of satellites for U.S. government constellations, be it the the Space Development Agency’s Proliferated Warfighter Space Architecture transport layer, or the recently disclosed Starshield constellation. The actual revenue associated with these projects is unknown, but considering the high value of the Starshield contract, it is reasonable to assume that these projects may yield $500 million and upwards in 2023 and 2024. 

HLS program. SpaceX is also involved in the NASA Artemis program as a supplier of the HLS (Human Landing System) to land US astronauts on the moon. This was a $2.9 billion contract, further topped by a second $1.15 billion contract. Payments for these contracts are linked to milestones, most of them seemingly tied to Starship achievements. It is likely that SpaceX already recognized large revenues from these contracts, maybe $700 million to $800 million a year. 

At Eurospace, we have designed our own SpaceX financial model as a tool to understand the intertwined dynamics of SpaceX revenues and expenditures, with a variety of underlying assumptions regarding costs and capex. 

One of these assumptions considers the actual costs of Falcon 9 launches, whose unique features of reusability, growing booster reuse rate (over 20 flights per booster), fairings recovery and reuse and acceleration of launch cadence (supported by rapidly decreasing refurbish times) mechanically drive average launch costs down. 

In 2021, Eurospace had modeled Falcon 9 financial assumptions and came to the conclusion that the reusability model generates large, and growing profits, at scale, after the seventh launch of the year (paid for by external customers). 

In 2024, the reusability model is vastly improved, and many analysts believe that the possibility that the full cost of each Falcon 9 launch —– including workforce for transport, refurbishment, assembly and operations, depreciation and amortization on facilities (launch sites, factories, test benches) and reusable items (fairings and boosters) — is currently positioned below $30 million. While some analysts make implicit assumptions that the full cost of Falcon 9 could be as low as $20 million per launch, Eurospace prefers to stay with the more conservative assumption of $28 million per launch. 

It is interesting to note that Falcon launch cost assumptions are the main driver of any analysis of SpaceX finances, and are the cornerstone to achieving profitability with the Starlink service. 

Indeed, when modeling the financials of Starlink with various (favorable to less favorable) assumptions on Capex and Opex, it appears that the single largest cost driver for Starlink is the space infrastructure Capex and related Depreciation and Amortization (D&A). 

In the Eurospace model, the assumption set leading to Starlink profitability in 2023 is intimately tied to the assumption of a Starlink Capex under $50 million per launch (satellites included), implying a full cost of a Falcon 9 launch below $30 million. For instance, if the Starlink Capex per launch were only as high as $60 million per launch, total Capex by end 2024 would be $24 billion. Such Capex would imply unsustainable D&A and vastly negative EBIT for the Starlink system alone. If this was the case, SpaceX would probably have had to secure more equity or debt financing to bridge the financial gap. This reverse reasoning is the strongest hint that indeed, for its Starlink system, SpaceX has achieved a capital efficiency never achieved by commercial constellations before, and 70% of that capital efficiency hinges on the low cost of Falcon 9 launches that SpaceX is charging to Starlink. 

The corollary to this assumption is that whenever Falcon is launched for external customers, including U.S. government and commercial or export customers, it is priced much higher than its actual cost, yielding a significant net profit as high as $30 million or more for a low-priced commercial launch, and up to $50 million to $60 million (and more?) when it is a governmental agency. In 2023, SpaceX undertook 33 launches for external customers, probably yielding a gross profit in excess of $1 billion. 

Interestingly, this thesis also goes towards the narrative of SpaceX profitability, since profit on launch will subsidize the Starlink system, covering the cost of maybe 40 to 50 Starlink launches (satellites not included). 

Eventually the thesis of less than $30 million per launch is a requirement for the narrative of Starlink breakeven or profitability statuses to be sustained. Moreover, it is the only thesis where this narrative can be sustained: if Falcon 9 is more expensive than that, there is no realistic set of assumptions where the SpaceX financial model does not go bankrupt. Or, to put it differently, if the Falcon 9 full cost significantly exceeds $20 million to $30 million, then either SpaceX is drowning in debt, or SpaceX has or had much higher revenues than what is currently estimated. 

As many commentators regularly note, SpaceX has managed to significantly lower the cost of access to space, but it is not passing these savings to its customers. Considering the dire situation of the competition (Arianespace and ULA mainly) SpaceX has actually no incentive to offer lower prices. 

In a recent article, some SpaceX competitors voiced their concern that SpaceX may be selling below costs to undercut them, and suggest that this is unfair. The fact that these competitors, notably Rocket Lab, exhibit much higher list prices per kg to orbit than SpaceX (upwards of $20,000per kilogram for Rocket lab, under $6,000 per kilogram for SpaceX), and still fail to turn a profit (as shown in the latest Rocket Lab financial statements), gives to their statements a strange resonance. Would it be possible, as they suggest, that SpaceX is selling launches at loss just to keep the competition at bay? If this was true, and due to the extremely high cadence of Starlink launches, it would mean that SpaceX is drowning in debt, incapable of sustaining the Capex drain of launching Falcons at more than $45 million to $60 million a pop. 

Of course, since SpaceX is private, nobody can prove or disprove that SpaceX is making a hefty profit on its Falcon launch business. But, as we discussed above, the assumption that Falcon 9’s full cost per launch is significantly higher than $30 million does not stand in front of the incredible launch cadence driven by the Starlink deployment. If we assume that each Falcon 9 actually costs SpaceX “as little” as $50 million, this would imply that SpaceX will have spent $10 billion and upwards on Starlink launch alone (plus a probable $6 billion to $8 billion for the satellites) by the end of 2024. We believe that this figure is not compatible with other financial facts about SpaceX, and would not support the affirmation that Starlink was breaking even on cash nor was profitable at some point of 2023. 

The most probable and logical assumption then is that by achieving reusability, fast turnaround and especially the very high cadence (thanks to the huge “demand” created by Starlink), SpaceX has unlocked a new economic model for space launch. This also means that SpaceX clearly has margins to undercut the competition while still making a profit on each launch (including the Transporter and Bandwagon mission and their relatively low fill ratio). 

Another interesting consequence of this assumption is that as long as SpaceX is the only launch operator able to benefit from low launch Capex for its constellation business, it will always have better financial prospects than its competition. 

For instance, let us consider Kuiper launch plans described as “the largest commercial launch deal ever” in which “Amazon is purchasing up to 83 launches from Arianespace, Blue Origin and United Launch Alliance.” At current market prices, it is assumed that the launch Capex from these 83 launches could be as high as $6 billion to $10 billion, deploying only half as many satellites as Starlink already did by mid-2024. If Kuiper had access to the same launch deal as Starlink, its launch Capex could be less than half of what it is estimated here. Moreover, without the Starlink low launch Capex, Kuiper service will suffer from a competitive disadvantage, and may never reach breakeven of profitability, which it could if SpaceX launch prices were not as high. 

Now that Kuiper recently negotiated a few launches with SpaceX, and may procure more in the future, the question of fair pricing may be worth raising. 

The very low costs of Falcon launches could also raise some questions from U.S. government customers. NASA and DoD regularly foot very expensive launch bills to launch with Falcon. Considering how much SpaceX’s successful development path has been tied to NASA and other U.S. government business, it may be reasonable for these customers to start wondering whether they have the right bang for the buck, and consider auditing SpaceX and determine whether the price offers made by SpaceX are fair or whether they include an excessive markup. 

We are really curious to see how this eventually unfolds, because it is very hard for the proponents and supporters of the newspace, Elon Musk included, to tout the dramatic reduction of launch costs as the key to unlock the “new space economy”, without ever seeing these cost reductions passed to the customers, and eventually unlock the benefits that “low-cost access to space” are supposed to bring to humanity. 

Pierre Lionnet is the Research and Managing Director at Eurospace, the trade organization of the European space industry). He labels himself a space economist, being an economist by training and being professionally involved in the analysis of space markets, space industry supply chains and space technology and innovation trends for the past 30 years. This op-ed reflects his own views, and does not constitute a formal position of Eurospace.