SpaceX: A Deep-Dive Research Report

Hunting ten-year five-baggers among great growth stocks — pressing the upside question: "Can it get much bigger?"

• QHow large is its market ceiling? Is it expanding an existing market, or creating an entirely new one?7/10

  A

  Conclusion: the ceiling is high enough, but “expanding an existing market” and “creating a new market” are happening in stages. Launch is essentially turning an existing market into an oligopoly. Starlink is expanding satellite broadband from a fringe test field into a new global connectivity market. The truly new markets created from scratch, such as orbital AI compute and the Earth-Moon economy, are still long-dated options today and have not yet monetized. This has to be assessed segment by segment. A vague “the TAM is enormous” answer is not good enough.

  Start with launch. This is taking share in an existing market and reshaping that market into an oligopoly. The report notes that among small satellites launched globally in 2020-2024, Falcon 9 alone carried 69%. In 2024, U.S. providers accounted for about 60% of global orbital launches, and about 70% were commercial launches. SpaceX is not staking out land in a new market. It has rewritten an old market, once run by national champions with low launch frequency and high prices, into an oligopolistic structure led by itself through reusability and high launch cadence. This market is not especially large on its own. Space segment revenue in 2025 was only about $4.1 billion, and the ceiling is constrained by total global launch demand.

  The real lift to the ceiling comes from Starlink. It is closer to taking the small “satellite broadband” market and expanding it into the much larger “global connectivity” market. The report and the S-1 show that by the end of March 2026, Starlink had about 10.3 million subscribers, coverage across 164 countries and regions, and about 9,600 satellites in orbit. Connectivity revenue in 2025 was about $11.4 billion. Satellite internet used to be a narrowband or geostationary niche business like Viasat and Iridium. Starlink has turned it into a horizontal network that can enter home broadband, enterprise leased lines, aviation, maritime, government, and direct-to-cell. This is not simply making the old market bigger. It is using the boundary of the old market, satellite communications, to attack the core territory of terrestrial communications. That includes both substitution, by taking customers from terrestrial operators, and new demand, from the 3 billion people on Earth with no or weak connectivity.

  As for “creating an entirely new market,” the report honestly places that in the long-dated option bucket, and it is basically not monetized today. Earth-Moon transport after full Starship commercialization, orbital manufacturing, and the orbital AI infrastructure imagined after xAI is folded in are the true “build a market from zero” pieces. But the report repeatedly stresses that these are “highly front-loaded, extremely volatile super-tech options.” The AI segment was still losing about $6.4 billion in 2025. During Starship V3’s 12th test flight on May 22, 2026, the booster segment still failed and the FAA ordered an investigation.

  So under Baillie Gifford’s “market ceiling” question, SpaceX’s addressable space in the connectivity layer is genuinely very large and still expanding, which supports a growth-stock narrative. But its current valuation of about $1.75 trillion is prepaying for a bundle of three things: an already enlarged launch oligopoly, an expanding global connectivity market, and space/AI markets that have not yet been born. The first two ceilings are real. The third is imagined. That is exactly why it is both attractive and dangerous.

  Jun 11, 2026
• QCan its revenue at least double over the next five years? Will growth mainly be driven by volume, price, or new businesses?6/10

  A

  Conclusion: revenue is very likely to double within five years, from about $18.7 billion in 2025 to more than $37.0 billion. The bar is not especially high. Growth will mainly be driven by Starlink “volume,” namely subscriber growth and expansion into more use cases. Price, or ARPU, is more likely to fall than rise. New businesses, such as Starship commercialization and AI contracts, are upside optionality rather than necessary conditions for doubling. But “can double” and “can make money after doubling and support the valuation” are different questions.

  Start with the base and feasibility. The report and the S-1 show consolidated revenue of about $18.67 billion in 2025, versus $14.02 billion in 2024, already above 33% growth in a single year. First-quarter 2026 revenue was $4.69 billion. For a company whose revenue is still growing at 30%+ and whose base has already been lifted by consolidating xAI, only about a 15% five-year CAGR is needed to double. Given the current scale and backlog, that is a relatively conservative target. The report’s “Neutral/Optimistic” scenarios also imply sustained high growth at Starlink.

  The main growth engine is Starlink “volume”. By the end of March 2026, it had about 10.3 million users and coverage in 164 countries, while the globally addressable population with no or weak connectivity is measured in billions. At the same time, aviation, maritime, enterprise leased lines, government, and direct-to-cell are all reusing the same constellation across more scenarios. The report judges that this business has “more installed terminals, broader coverage, and lower unit cost.” Connectivity revenue in 2025 was about $11.4 billion, or 61% of consolidated revenue, making it the main driver of a doubling.

  But “price” is more likely to be a drag than a help. In its pre-mortem, the report lays out a specific scenario in which “Starlink growth quality deteriorates”: users keep growing, but consumer broadband ARPU falls faster than high-value use cases such as aviation and government ramp, pulling the connectivity segment’s operating margin down from the current 30%+ level. In other words, volume rises while unit price is diluted by the consumer side, and the offset has to come from mix improvement in higher-value use cases such as aviation, maritime, government, and direct-to-cell. Whether that offset works is the key to margins, not revenue scale.

  New businesses, Starship and AI, are upside elasticity, not prerequisites for doubling. If Starship crosses the commercial reusability threshold, it will sharply reduce launch costs and open incremental Earth-Moon transport revenue. AI segment revenue was about $3.2 billion in 2025, and if xAI can deliver high-quality contracts, it could also contribute scale. But today one of these is still in test flight, with the V3 booster failing on May 22, and the other is posting huge losses, with AI segment operating loss of about $6.4 billion in 2025. Counting them as necessary conditions for a revenue doubling would be dishonest. They are better understood as long-term variables that decide whether growth can accelerate after the doubling and whether the valuation can be justified.

  The short version: the probability of a revenue doubling is high, and the main axis is Starlink volume growth. That satisfies Baillie Gifford’s growth requirement. But the report’s real warning is that the “quality” of the doubling, meaning margins not being eaten by ARPU pressure and capital expenditure, is much harder than the “shape” of the doubling, and the about $1.75 trillion valuation is buying the latter.

  Jun 11, 2026
• QAfter five years, what will take over as the next growth engine? Does this “second curve” exist today?5/10

  A

  Conclusion: SpaceX’s “second-curve” candidates do exist today, and there is more than one: direct-to-cell/government national-security networks in the near term, Starship commercialization in the medium term, and xAI orbital AI infrastructure in the long term. But their certainty declines as their cash burn rises. What is visible today is “investment,” not yet a stable “cash-flow handoff.” The problem is not the absence of a second curve. The problem is that SpaceX is betting on three at once, and none has yet proven it can fund itself.

  First define the main curve: today’s real profit engine is Starlink, the connectivity segment, with operating profit of about $4.42 billion in 2025 and another about $1.19 billion in the first quarter of 2026. When Baillie Gifford asks what takes over after five years, the real question is: once consumer broadband penetration matures and ARPU falls, what extends growth?

  Second-curve candidate 1, nearest and highest certainty: direct-to-cell plus government/national-security networks. This is a natural extension of Starlink’s main curve, not a brand-new business. The report notes that direct-to-cell has received FCC approval, that there is additional authorization for 7,500 second-generation Starlink satellites, and that EchoStar’s approved 65MHz spectrum transaction gives it more cards in hybrid cellular/satellite. On the government side, SpaceX won 28 missions worth more than $5.9 billion under NSSL Phase 3 Lane 2, on top of Starshield and potential missile-defense networks. The strength of this curve is that it reuses the existing constellation in higher-ARPU scenarios. The weakness is that it is still attached to Starlink and does not quite count as an independent second growth pole.

  Second-curve candidate 2, medium term and medium certainty: Starship commercialization. The report repeatedly emphasizes that if Starship achieves high-frequency recovery and orbital refueling, it will further reduce costs and open a new Earth-Moon transport profit pool. That would be a genuinely independent new curve. But today it remains in engineering validation: the V3 12th test flight on May 22, 2026 deployed 20 simulators plus 2 real satellites and completed a controlled splashdown, but the booster still failed and the FAA ordered an investigation. The inflection from story to cash flow has not arrived.

  Second-curve candidate 3, farthest, lowest certainty, and most cash-intensive: xAI / orbital AI infrastructure. After xAI was folded in during February, the company pushed the narrative toward a “space + connectivity + AI” super-platform. But the AI segment had only about $3.2 billion of revenue and about $6.4 billion of operating loss in 2025. In the first quarter of 2026, it lost another $2.47 billion and accounted for 76% of the quarter’s $10.1 billion of capital expenditure. This “curve” is a pure investment item today, and even its direction is still being questioned by firms such as Morningstar, which assigned the AI business only about $170 billion of probability-weighted value and treated it as the main reason for lowering its rating.

  The honest Baillie Gifford view: second curves do exist today, and the reserve depth is far greater than at most peers. Rocket Lab and Iridium do not have this kind of depth. That is a positive. But handoff certainty is sharply uneven. The near-term curve, government/direct-to-cell, is more certain but has limited incremental scale. The far-end curves, Starship and AI, have huge imagination value but have not yet delivered cash flow. The report therefore characterizes SpaceX as “undergoing valuation reshaping, with a larger component of future prepayment”: the market is already pricing it at about $1.75 trillion as if all three curves successfully take over, while the only one repeatedly proven is Starlink.

  Jun 11, 2026
• QWhat is its core competitive advantage? Will this moat widen or narrow over the next three to five years?7/10

  A

  Conclusion: the moat is real and unusually strong. Four layers reinforce one another: reusable cost curve, vertical integration, network scale, and government/national-security trust. Over the next three to five years it is more likely to widen than narrow. But there is one structural crack, concentrated governance, where the person is both the moat and the risk. That cannot be repaired in the short term and is the only real hazard inside this wide moat. This is one of the few assets where the quality of the business itself does not require much debate.

  The first moat: cost advantage and learning curve from reusability. The report notes that a single Falcon 9 first-stage booster has flown its 30th mission, and that Falcon 9 handled 69% of global small-satellite launches in 2020-2024. This is not a one-off technical stunt. It turns heavy assets that used to be discarded after one use into a “fleet” that cycles dozens of times. Others run projects; SpaceX runs a liner service. Each additional flight makes the cost curve harder to catch, a classic widening-through-use moat.

  The second: vertical integration. Rockets, satellites, ground stations, terminals, launch sites, and government delivery are all inside its own system, which improves efficiency and speeds feedback loops. For the same capital expenditure, it can capture more synergy than fragmented competitors. In its peer comparison, the report emphasizes that this is exactly why SpaceX “gets more out of the same capex than others.”

  The third: network scale and service density. By the end of March 2026, Starlink had about 10.3 million users, coverage in 164 countries, and about 9,600 satellites in orbit. Satellite internet has powerful economies of scale: more installed terminals, lower unit cost, denser coverage, and a harder time for competitors to replicate the same constellation. Spectrum and orbital slots are scarce resources. This moat widens with each batch of satellites launched.

  The fourth: government and national-security trust. It began with NASA’s $1.6 billion ISS resupply contract in 2008 and extended to commercial crew, the $2.89 billion HLS lunar landing contract, 28 missions worth more than $5.9 billion under NSSL Phase 3 Lane 2, and Starshield. This kind of trust is built from more than a decade of delivery records. Marketing cannot buy it, and new entrants can hardly replicate it within three to five years.

  Will it widen or narrow over the next three to five years? My judgment is that it widens: reusability and launch cadence are still pulling ahead, Starship V3 has begun higher-frequency Starlink deployment, network scale will keep growing with expansion authorization, including another 7,500 second-generation constellation satellites, and government orders are increasing. The most realistic competitor, Amazon Kuiper, has abundant resources but started late and remains an order of magnitude behind in scale. The report judges that it will be hard for Kuiper to shake SpaceX on the main battlefield in the short term.

  But the one structural crack has to be stated plainly: governance. The report calls this the “most obvious weakness.” After the IPO, Musk retains 85.1% voting control, with 10 votes per Class B share, plus mandatory arbitration, limits on litigation venue, and even provisions that “protect Musk from being fired by anyone other than himself.” This is a double-edged structure where the person is the moat and also the governance risk. In execution, Musk is the biggest asset. In governance, he is the biggest source of discount. And this crack cannot be naturally repaired by one financial report.

  Bottom line: the business moat is wide and widening, fully meeting Baillie Gifford’s standard for a “great company.” The only thing pulling the score down is not the moat itself, but the highly concentrated governance structure sitting over it. It does not determine whether the company can execute, but it will amplify the valuation discount in a headwind.

  Jun 11, 2026
• QIf its core business is disrupted, does it have the genes for self-reinvention? How does it treat errors and bad news?6/10

  A

  Conclusion: SpaceX has the strongest “self-reinvention genes” in its peer group. Its entire history is a record of repeatedly jumping to new business frontiers and treating engineering failure as iteration fuel. In dealing with errors and bad news, its engineering culture is quite healthy: public test flights, investigate after explosions, then fly again. But the same genes become a risk when magnified through finance and governance. A culture that treats failure as iteration fuel, combined with 85.1% absolute control, means correction depends entirely on one person and outside shareholders have almost no balancing lever. Reinvention capacity is extremely strong. Checks and balances are extremely weak.

  Start with the “self-reinvention genes,” its clearest strength. The report’s longitudinal history is itself a chain of proactive reinventions: from rockets to spacecraft, from expendable to reusable, from a launch company to the Starlink constellation network, and then in 2026 to a “space + connectivity + AI” hybrid after xAI is folded in. The report states clearly that the company “is not just going deeper on one track, but repeatedly proving one capability across launch, crewed flight, defense, and broadband.” A company that can repeatedly jump to new frontiers over 20 years and execute each time is exactly the kind of “regenerate itself even if the core is disrupted” organism that Baillie Gifford looks for.

  Now consider “how it treats errors and bad news.” On the engineering side, it is textbook healthy. Starship’s development philosophy is public, high-frequency, and tolerant of explosive iteration: the V3 12th test flight on May 22, 2026 successfully deployed 20 simulators plus 2 real satellites and completed a controlled splashdown, but the booster failed and the FAA immediately ordered an investigation. SpaceX’s normal mode is to treat this kind of failure as data, investigate the defect, and keep flying. The report also notes that Falcon 9 was briefly grounded in February due to a second-stage issue, showing that even for its most mature system, it accepts the discipline of “ground, investigate, return to flight.” This culture of not hiding failure and feeding failure into iteration is a real advantage.

  But the hidden premise of this Baillie Gifford question, whether the company can reinvent itself if the core business is disrupted, requires the other side to be laid out: the steering wheel for reinvention is held by one person. The report lists governance as the most obvious weakness. After the IPO, Musk holds 85.1% voting control and serves as CEO/CTO/chairman. The company will use its “controlled company” status to be exempt from majority-independent-board requirements and has mandatory arbitration and venue limits. That means when a painful turn is truly needed, decisions can be made extremely fast, which is the benefit. But if the person at the wheel judges wrongly, outside shareholders have almost no corrective tools, which is the downside. Combining the engineering gene of “tolerate failure and iterate fast” with a governance structure lacking external checks makes correction speed depend on Musk personally. In good times it is an asset; in bad times it is a single point of risk.

  There is also a financial layer to “bad-news handling” that deserves caution. The largest current “error exposure” is the AI segment’s huge loss, with operating loss of about $6.4 billion in 2025 and another $2.47 billion loss in the first quarter. The company’s chosen “reinvention” is to keep using Starlink cash flow to feed the most distant story. Whether that is active capital allocation or using a good business to cover a bad one while refusing to stop loss, the report admits that “both interpretations have factual support.” This is what Baillie Gifford should watch: the reinvention gene is genuinely strong, but it may also be used to avoid admitting that a long-dated bet should be cut.

  Bottom line: SpaceX’s engineering and strategic capacity for self-reinvention is one of the qualities that stands up best to Baillie Gifford scrutiny. The deduction is not about whether it can reinvent itself. It is about who decides the reinvention, and whether outside shareholders can stop it if it goes off course. The answer is: almost certainly not.

  Jun 11, 2026
• QDoes management, especially the founder, have a long-term view and deep alignment with the company? Is it willing to sacrifice current profits for five to ten years from now?7/10

  A

  Conclusion: on “long-term view” and “willingness to sacrifice current profits for five to ten years from now,” SpaceX is the kind of company Baillie Gifford would score at the top. Musk’s interests are deeply tied to the company, with 85.1% voting control and massive equity ownership. The mission is explicitly measured over decades, making humanity a multiplanetary species, and the company is taking large current losses to fund long-term bets. But the same facts sharply reduce governance friendliness: his alignment comes through absolute control, so outside shareholders are in the same boat as the founder but have no steering wheel. Long-termism is full-score quality. Shareholder friendliness is extremely low. Both come from the same fact.

  Start with “long-term view,” which is exactly the ideal type in the Baillie Gifford framework. The report opens with the company’s mission to “make humanity a multiplanetary species” and points out that the entire development history is a 20-year path of turning that slogan into “first change the cost curve, then change the market boundary.” Musk’s decision horizon has never been quarterly. It is decade-scale: taking the $2.89 billion HLS lunar contract in 2021 tied Starship into the national deep-space program, and Starlink cash flow continues to be invested into the farthest-out Starship and AI bets. This willingness to subordinate current profits to a long-term vision is exactly what Baillie Gifford’s “will it sacrifice now for five to ten years from now” question is meant to find.

  The willingness to sacrifice current profits for the long term is proven with real money. The report and the S-1 show that the company had $791 million of profit in 2024, then deliberately swung to a net loss of about $4.94 billion in 2025. First-quarter 2026 capital expenditure reached $10.1 billion, 76% of which went into AI. This is not passive loss-making. It is the deliberate use of already-profitable Starlink cash flow to feed Starship and xAI. It is one of the most extreme real-world examples of sacrificing current profit for five to ten years from now.

  On interest alignment, it is hard to find anything deeper. According to the S-1, Musk holds 8.495 million Class A shares (12.3%) + 557 million Class B shares (93.6%), totaling 85.1% voting control. His personal wealth is highly coupled to the company’s fate. That is the founder “putting his net worth on the table” that Baillie Gifford likes to see.

  But this question cannot be answered only in its first half. The form of that alignment has to be laid bare, and that is exactly where the deduction lies. The report lists governance as the most obvious weakness: 10 votes per Class B share, use of “controlled company” status after the IPO to avoid majority-independent-board requirements, mandatory arbitration, venue limits, and reportedly even arrangements that “protect Musk from being fired by anyone other than himself”. In other words, founder-company alignment is extremely deep, which is good, but outside shareholders are “tied to the same boat with no steering wheel.” The report calls this “the person is the moat, and also the governance risk.” Morningstar’s governance team also explicitly lists share structure and board composition as core concerns for institutional investors.

  One further potential conflict of interest must also be stated. Musk simultaneously runs Tesla and X, while xAI was his own asset before being folded into SpaceX in February. Putting an AI company he controls into SpaceX, then asking the public market to pay for that “hybrid” at about $1.75 trillion, inevitably raises questions about related-party transactions and divided attention. The report also identifies xAI consolidation as a source of a blurred valuation anchor.

  Bottom line: if this Baillie Gifford question asks only whether the vision is long enough, whether management is willing to sacrifice today for the long term, and whether the founder’s wealth is tied to the company, SpaceX is a model “yes.” But it has turned “alignment” into “absolute control + no checks.” The same person therefore scores near the top on long-term trust and near the bottom on friendliness to minority shareholders. That is the root reason the report gives “management credibility” only a “Medium” and assigns the overall rating “Watch.”

  Jun 11, 2026
• QIf it disappeared tomorrow, how much would customers miss it? Is its growth sustainable and not dependent on harming society or regulators?7/10

  A

  Conclusion: indispensability is extremely high. If SpaceX disappeared tomorrow, U.S. government access to space, NASA’s deep-space and crewed programs, more than 10 million Starlink users, and even most global small-satellite deployments would instantly lose a pillar, with almost no equivalent substitute. Its growth model overall does “not depend on harming society,” because it provides real connectivity and launch services and fills gaps for unconnected populations. But it is highly dependent on regulatory permission and has sustainability concerns around spectrum occupation, orbital debris, and governance friendliness. The indispensability half is almost full-score. The social/regulatory sustainability half is acceptable but qualified.

  Start with “indispensability,” its hardest edge. The report notes that by the end of March 2026, SpaceX had completed about 620 orbital launches with a mission success rate above 99%, and that from 2020 to 2024 it alone handled 69% of global small-satellite launches. On the government side, it is deeply embedded in U.S. national space capability, from NASA commercial crew and the $2.89 billion HLS lunar contract to 28 national-security missions worth more than $5.9 billion under NSSL Phase 3 Lane 2. If it disappeared, the combined near-term capacity of ULA, Blue Origin, and Rocket Lab could not fill that gap. This position, where both the government and an entire industry cannot do without you, is one of the strongest answers to Baillie Gifford’s “how much would customers miss it” question.

  The Starlink side is just as hard: 10.3 million users across 164 countries, many in remote areas, maritime, aviation, emergency response, and battlefield communications where terrestrial networks do not reach. For these users, Starlink is not a “cheaper replaceable option.” It is often the “only usable connection.” The degree to which they would miss it is extremely high.

  The hidden double premise in this Baillie Gifford question, beyond “indispensable,” is whether growth is sustainable and does not depend on harming society or regulation. That has to be answered honestly in two layers.

  Layer one, “does it harm society”: overall, positive or neutral. Its core growth comes from providing real value: lowering launch costs, connecting populations with no or weak networks, and supporting national security and scientific research. It is not growing through regulatory arbitrage or by harming users. On this point, it is cleaner than many businesses whose growth depends on squeezing externalities.

  Layer two, “does it depend on regulation / are there sustainability concerns”: there are real reservations here and they should not be papered over. First, its growth is deeply dependent on regulatory clearance: launch licenses from the FAA, spectrum licenses from the FCC, orbital debris rules, and national-security reviews are hard constraints. The report lists “regulatory/governance controversy escalation” as a standalone risk, and the FAA’s immediate investigation order after the V3 booster failure is a live example. Second, nearly 10,000 satellites in orbit, plus authorization for another 7,500 second-generation satellites, raise long-term questions around orbital congestion, debris, and interference with astronomical observation. Third, on governance, 85.1% voting control + mandatory arbitration + limits on shareholder litigation are treated by the report as structural discounts that are invisible in tailwinds and priced in headwinds. From the broader sustainability angle of “is it friendly to minority shareholders,” that is also a negative.

  Bottom line: under Baillie Gifford’s standard, on the “how much would customers miss it if it disappeared tomorrow” half, SpaceX gives one of the strongest answers in the industry. Government, research, and more than 10 million users are highly dependent on it, and it is hard to replace. On the “is growth sustainable and not dependent on harming society or regulators” half, it is not growing by harming society, so it passes, but growth is tightly constrained by regulatory permission and carries two long-term externality debates, orbital environment and governance. So the overall answer is positive, but not a flawless full score.

  Jun 11, 2026
• QWhat are the unit economics of this business, including gross margin and incremental returns? Do they improve or deteriorate with scale? Where does the money it earns go?5/10

  A

  Conclusion: unit economics are a “one good, two difficult” mix. Starlink, the connectivity segment, has excellent unit economics that improve with scale, with deferred investment and recursive monetization, and an operating margin near 39% in 2025. But consolidated unit economics are severely diluted by the capital intensity of the Space segment and the huge losses in AI. Almost all the money earned, and more than all of it, is being spent on AI and Starship, turning the company from “expanding with profits” into “using profitable segments to fund loss-making segments, then relying on capital markets for more fuel.” By segment it looks like an honors student. Consolidated, it looks like a cash-burning machine. The key is whether Starlink can earn money faster than AI/Starship can spend it.

  The best piece, Starlink or the connectivity segment, has unit economics that improve with scale. In 2025, connectivity revenue was about $11.4 billion and operating profit about $4.42 billion, implying an operating margin of about 39% (44.23÷113.87), with growth of about 49.8% from the prior year and EBITDA growth of 86.2%. The report says the business has “more installed terminals, broader coverage, lower unit cost, and better marginal profit.” In substance, it is recursive monetization in a platform/telecom model. That is exactly the kind of “unit economics improve with scale” business Baillie Gifford favors. Once the same network supports aviation, maritime, government, and direct-to-cell, ARPU mix can also shift toward higher-value scenarios.

  The hardest piece, the AI segment, has negative unit economics and is consuming cash. In 2025, the AI segment had about $3.2 billion of revenue but about $6.4 billion of operating loss, and then lost another $2.47 billion in the first quarter of 2026. There are no positive unit economics in this segment today. It is a pure investment item.

  The middle piece, the Space segment, is a project/platform hybrid with nonlinear unit economics. It will not improve linearly like SaaS, because it has to sustain heavy spending on manufacturing, testing, and launch sites over the long term. It posted a small operating loss of about $660 million in 2025. But reusability and higher launch density can create strong operating leverage. A single Falcon 9 booster has flown 30 times, spreading fixed assets across dozens of missions and making the asset cheaper to use with each flight.

  Consolidated, unit economics are severely diluted. In 2025, consolidated revenue was about $18.67 billion, but operating loss was about $2.59 billion and net loss was about $4.94 billion, versus $791 million of profit in 2024. In other words, the $4.4 billion that Starlink earned was not enough after being consumed by Space losses plus the AI $6.4 billion hole. High-quality incremental returns exist inside the connectivity segment, but on the consolidated statement they are offset by bad businesses.

  Where does the money earned go? Almost all of it, and more, is being poured into AI and Starship. Total capital expenditure in 2025 was about $20.7 billion, 61% of it AI. First-quarter 2026 capital expenditure was $10.1 billion, still 76% AI. Note that first-quarter capital expenditure of $10.1 billion far exceeded that quarter’s $4.69 billion of total revenue. That means the money is coming not only from Starlink profit, but also from debt and financing. The report says bluntly that the company has moved from “using profit to support expansion” to “using profitable segments to fund loss-making segments while continuing to use capital markets for fuel,” and notes that a $20 billion bridge loan reduced debt in early March to about $20.07 billion. The S-1 quarter-end figure is higher, and the report uses the quarter-end basis to judge that leverage is “not light.”

  Baillie Gifford bottom line: the “most profitable segment” of this business has excellent unit economics that improve with scale, and it clearly qualifies. But SpaceX is putting the money it earns, and even more than it earns, almost entirely into AI that does not make money today and Starship that is still in test flight. So the real Baillie Gifford question is not “are Starlink economics good” (they are), but “is capital allocation discipline good enough.” The report repeatedly stresses that the most important metric to track is whether Starlink can keep earning money faster than AI and Starship can spend it. If that does not hold over time, high-quality unit economics will not translate into shareholder returns.

  Jun 11, 2026
• QWhat conditions must hold simultaneously for it to rise fivefold over ten years? Are those conditions realistic? What expectations are implied in today’s share price?2/10

  A

  Conclusion: starting from about $1.75 trillion, a fivefold gain over ten years would require a market value of about $8.75 trillion. That would require Starlink to sustain high growth without stalling, at least one and preferably both of the two long-dated curves, Starship and AI, to turn from story into large-scale cash flow, margins and capital expenditure to inflect at the same time, and the market to maintain a top-tier premium multiple for a long time. Multiple demanding conditions must all hold simultaneously. Today’s issue price already embeds extremely front-loaded expectations: Morningstar’s estimated fair value is only about $780 billion, less than half the target issue value, meaning the market is already prepaying a large amount of “future not yet delivered.” A fivefold return is not impossible, but it requires all bets to land at once, while the starting valuation has already prepaid much of the future.

  First, lay out the arithmetic of “fivefold.” The starting point is the IPO target of about $1.75 trillion, with $135/share, 555.6 million shares issued, and about $75 billion raised. Fivefold over ten years means about $8.75 trillion, far above Microsoft today and near an unprecedented market-value scale. Supporting that cannot be done with only about $18.7 billion of 2025 revenue, which already implies about 94 times sales. Revenue and profit would need an order-of-magnitude leap.

  “What conditions must hold simultaneously?” Under the report’s logic, at least these five must all happen:

  1. Starlink must not stall. It must keep growing rapidly for ten years from a base of 10.3 million users, while margins are not crushed by declining consumer ARPU. The current connectivity segment operating margin is about 39%. The report’s second pre-mortem scenario is exactly “users keep rising, but ARPU collapses and the narrative collapses.”
  2. Starship must move from engineering validation to commercial validation. It must achieve high-frequency recovery, orbital refueling, and scaled commercial payloads, opening a new Earth-Moon transport profit pool. Today it has just had a booster failure in the May 22 V3 test flight and is under FAA investigation.
  3. The AI segment must turn from huge losses into high-quality cash flow. Current 2025 operating loss is about $6.4 billion. It must become an infrastructure platform that produces contract cash flow. This is the piece Morningstar sees as the most uncertain and the basis for its rating cut.
  4. The capital expenditure/revenue ratio and debt curve must turn. First-quarter capital expenditure of $10.1 billion far exceeded revenue, and the company relied on a $20 billion bridge loan to keep going. It must move from cash burn to positive free cash flow.
  5. The market must maintain a top-tier premium multiple for a long time. Even if fundamentals deliver, capital markets must continue to assign a technology-platform valuation for ten straight years, while governance, including 85.1% voting control, remains a source of discount hanging over that premium.

  Are these conditions realistic? Looked at individually, Starlink high growth is credible, long-term Starship success is possible, and AI turning positive is possible. But the force of this Baillie Gifford question lies in “simultaneously.” Five highly uncertain things must almost all land. The joint probability is far lower than that of any single item. Even the report’s optimistic scenario gives reasonable equity value of only about $1.9 trillion, just +9% versus the issue price. The report’s own bull case is still orders of magnitude away from “fivefold,” showing that a ten-year fivefold return would require execution far beyond the report’s optimistic scenario.

  What expectations are implied in today’s share price? A large amount of future not yet delivered has already been prepaid. The clearest evidence is the gap with third-party valuation: Morningstar gives fair value of about $780 billion (about $63/share), a roughly 53% discount to the $135 issue price. Within that, DCF assigns about $611 billion of enterprise value to core launch + Starlink, and only about $170 billion to AI on a probability-weighted basis. In other words, the issue price is not pricing cash flow already landed. It is prepaying for a chain of unproven options: full Starship commercialization, large-scale monetization of direct-to-cell, orbital AI infrastructure, and more. The report’s wording is consistent: at $1.75 trillion, investors are “buying faith more than buying free cash flow already delivered.”

  Baillie Gifford bottom line: a fivefold rise over ten years is not physically impossible, but it requires “five demanding conditions to hold simultaneously + valuation multiples not to collapse for ten years,” a low joint-probability scenario. More importantly, the starting line has already prepaid a large slice of the future. When fair value estimates are only half the issue price, the margin of safety for a “fivefold” outcome is almost negative. That is the fundamental reason the report scores valuation attractiveness “Low,” sets the rating at “Watch,” and states that using $1.75 trillion as the buy-in baseline implies a conservative expected annualized return of -15% to -20%.

  Jun 11, 2026
• QWhy has the market not realized all this yet? Is it because investors do not understand it, look down on it, or cannot look far enough? What will become the “narrative inflection point”?2/10

  A

  Conclusion: this company is unusual because the market has not “failed to realize how good it is.” Quite the opposite: the market may have realized it too much, pricing good assets and bad assets as if they were the same quality in one bundle. If we must apply Baillie Gifford’s “do not understand / look down on / cannot look far enough” framework, the real underappreciated insight is not “how great the company is,” but “how different the certainty levels are across its assets.” The market can see Starlink making money, but can easily fail to see that AI/Starship realization is still far away. Narrative inflection points will appear in the first batch of post-listing financial reports, Starship commercialization milestones, and whether capital expenditure and debt curves turn. This is one of the rare cases where the market is not dismissive, but may be too hot.

  First correct the premise: for SpaceX, this Baillie Gifford question has to be asked in reverse. The usual Baillie Gifford script is “a great company is underestimated by the market, waiting for the perception gap to close.” But the report states clearly that there is relatively little dispute over whether SpaceX is an excellent company. The real disagreement is “whether it is still worth buying at the current valuation.” In other words, the market’s recognition that “it is excellent” is already ample and perhaps excessive: the IPO recorded about $150 billion of demand, the target valuation is about $1.75 trillion, and the price-to-sales ratio is about 94 times. This is not “being looked down on.” It is being chased aggressively.

  So what might the market truly have “not realized”? The layering of asset quality. The report’s core insight is that the market’s most likely misjudgment is “not that SpaceX cannot execute, but that assets with different levels of certainty are being bundled and valued as if they were the same quality.” Starlink’s profit quality, with 2025 connectivity segment operating profit of about $4.4 billion and margins near 39%, the Space segment’s capital intensity, and the AI segment’s loss profile, with a 2025 loss of about $6.4 billion, are fundamentally different. Once blended into a “super-tech platform,” the market can overestimate the ability of the good assets to cover the bad assets. In Baillie Gifford’s three-part framing, this is not “do not understand” or “look down on.” It is closer to “looking too far and pricing long-dated options as near-term certainty”. Clear-eyed voices already exist among professional institutions: Morningstar gives fair value of only about $780 billion, less than half the issue price. That shows the “perception gap” is precisely the disagreement between optimists and cautious investors over the probability that AI/Starship will deliver.

  The implied follow-up in this Baillie Gifford question, “what becomes the narrative inflection point,” is answered by the report with a clear observable list:

  • Inflection point 1: the first batch of quarterly reports after listing. This will be the public market’s first look at the “persistence” of Starlink/Space/AI rather than an IPO prospectus snapshot. If Starlink slows because of scale, or AI losses remain out of control, the narrative will immediately be repriced downward.
  • Inflection point 2: Starship commercialization milestones. Especially recovery, orbital refueling, and scaled commercial payloads. The May 22 V3 test-flight booster failure and FAA investigation order are reminders that the biggest upside narrative remains stuck in engineering validation, and any key success or failure can be an inflection point.
  • Inflection point 3: whether capital expenditure and debt curves slow. First-quarter capital expenditure of $10.1 billion far exceeded revenue, and the company relied on a $20 billion bridge loan to keep going. The report says this is “the key to deciding whether it is truly a growth stock or a high-valuation cash-burning stock.”

  One technical inflection point also needs to be added: lockup structure. This IPO does not have a standard 180-day lockup. It allows existing shareholders to sell 20% after the first quarterly report, then releases shares in batches at 70/90/105/120/135 days. When seller supply exceeds buyer demand in certain windows, that alone can become a sentiment inflection point, unrelated to fundamentals but able to amplify volatility.

  Baillie Gifford bottom line: for most companies, this question searches for “buried greatness.” For SpaceX, the honest answer runs the other way. The market already fully recognizes its greatness. The real perception gap lies in treating assets with different certainty as if they were the same quality. The narrative inflection point is not “when the market finally understands it is good,” but “which of these three gets confirmed by financial reports first: Starlink stalling, Starship getting stuck, or cash burn persisting.” This also matches the report’s final conclusion: first-class asset, excessive expectations, extremely weak governance, rating “Watch.”

  Jun 11, 2026