SpaceX: A Long-Term Owner's Perspective

Business comprehensibility score: 3/5. If you look only at Falcon, Dragon, Starlink, and Starshield, this company can be understood: it sells launch capacity, orbital logistics, communications connectivity, and security networks. The problem is that once you treat "the entity currently preparing to list" as the investment target, you must fold in xAI, AI data centers, and even space-based data centers, which turns it from "a comprehensible industrial infrastructure company" into "a hybrid wrapped in a grand narrative." Through a Buffett-style framework, the core business is understandable, but the combined whole is no longer simple and transparent enough.

If the stock market were to close for the next five years, I would be willing to hold the core SpaceX business at the right price. But I would not want to be passively locked in for the long term at the current valuation band of $1.25 to $1.75 trillion, carrying the uncertainty of xAI consolidation. This is not distrust of the company's quality; it is restraint regarding the entry price and the governance structure.

The Industry and Competitive Landscape

SpaceX does not operate in a single industry. It spans at least three: commercial launch, low-Earth-orbit satellite communications, and government/defense space infrastructure. The common feature of these three arenas is long-term rising demand, all with strong regulatory, capital, and technological barriers. Reuters reports that Falcon completed 165 launches in 2025, nearly "one launch every two days"; the FAA has approved raising Starbase's annual Starship orbital launch cap to 25. This capability is not merely about speed; it turns "reaching orbit" from a sporadic engineering task into something resembling an industrial production line.

On the competition: in launch, the main rivals are ULA, Blue Origin, and Rocket Lab; in low-orbit internet, primarily Amazon Leo and Eutelsat OneWeb; in direct-to-cell and more specialized satellite connectivity, AST SpaceMobile is also advancing quickly. Amazon's official materials show that Leo, formerly known as Kuiper, targets broader commercialization in 2026 and continues deploying satellites; Eutelsat/OneWeb currently operates more than 600 low-orbit satellites and is pushing into government, defense, aviation, and maritime markets; Blue Origin won a $3.4 billion NASA Artemis lunar lander contract and participates in national security launch. In short, SpaceX is not without competitors; it leads in most arenas, but does not hold a monopoly.

Long-term industry demand is broadly positive. Global demand for low-latency wide-area communications, resilient defense networks, remote sensing and reconnaissance, and government mission-critical launch will very likely persist for the long term, and geopolitics tends to reinforce it. The real question is not "is there demand" but "who can satisfy it at a low enough cost, with high enough reliability, and with high enough regulatory-compliance efficiency." On that question, SpaceX's core competitiveness is genuinely real.

But from a value-investing perspective, this still looks more like "an excellent company in a poor industry" than "a naturally asset-light, good industry." Spaceflight and satellite communications are inherently capital-intensive, policy-sensitive, and unforgiving of failure, with rapid technological substitution. SpaceX has turned an industry that historically delivered terrible returns into something close to a platform-style infrastructure business, which is remarkable. But that does not automatically mean any price is worth paying. Industry attractiveness score: 4/5; yet the industry's inherent capital intensity still lowers its purity within a Buffett framework.

Moat and Management

Moat Analysis

SpaceX's moat is strongest not in "brand premium" but in a systemic moat formed by the overlap of cost advantage, scale advantage, regulatory/credential advantage, and engineering-culture advantage. Falcon 9's official price is $74 million; Starship is officially positioned to put larger payloads into orbit at a lower marginal cost than the current Falcon. Falcon's high-frequency reuse, Starlink's internal demand, and vertically integrated satellite manufacturing and launch capabilities give SpaceX a flywheel that is extremely hard to replicate: internal demand feeds external supply, and external supply in turn spreads internal costs.

The evidence for this moat is highly concrete. SpaceX does not just run technology demonstrations; it has already converted this capability into commercial contracts and launch cadence: commercial crew and cargo, lunar landers, national security launch, ISS deorbit, the Starshield government security network, and the NRO satellite network all rest on the same manufacturing, launch, and operations backbone. In other words, even a competitor that can build a rocket may not be able to replicate this composite system of "launch capability + orbital operations + communications network + government certification."

Breaking the moat apart: Brand advantage: Yes, but it shows up more in recruiting top engineers, winning government trust, and customer preference, than in a consumer-brand premium. Cost advantage: Very strong, derived from reuse, cadence, manufacturing scale, and internal payloads. Scale advantage: Very strong; a cadence of 165 launches a year is very hard to replicate. Network effects: Starlink has some "the wider the coverage, the more data, the stronger the terminal ecosystem" network effect, but it is not as pure as a typical software platform. Switching costs: Medium-to-high for government, aviation, maritime, and remote enterprise customers; medium for ordinary consumers. Distribution advantage: Mainly access to the government/defense procurement system and the ability to land internationally. Licensing/regulatory barriers: Very strong, especially the multilayered credentials across the FAA, FCC, Space Force, and NASA. Data advantage: Some accumulation at the network-operations and space-mission-execution level, but not a traditional internet-style data monopoly. Corporate culture and operating capability: Extremely strong; this is the part I consider hardest to replicate. Capital allocation capability: Historically excellent, but markedly more aggressive over the past two years.

My judgment is this: the core moat is still widening, but the combined entity's "investability" has not widened in step. In other words, SpaceX's core business has grown stronger, yet the position of minority shareholders has not improved accordingly. To replicate this system, a competitor needs not only billions of dollars in capital, but also years of safety record, frequent mission experience, government access, launch-site infrastructure, supply-chain ramp capability, and large enough internal payload demand. Moat strength score: 4.5/5.

On inflation and recession: in theory, SpaceX has some pricing power, because launch supply remains a scarce resource and Starlink is an "irreplaceable connection" in many regions. But it is not a typical asset-light software service and cannot raise prices nearly painlessly the way Visa or Microsoft can. In a recession, its government/defense business and remote-connectivity business would be more stable, but Starlink terminal sales, some commercial launch demand, and the capital efficiency of high-risk R&D would all come under pressure. It is not a business that "easily sustains high margins even in a downturn." It is a business that "can still survive hard times, but with volatile returns on capital."

Management and Capital Allocation

If we evaluate only the ability to "turn complex technology into commercial reality," the SpaceX led by Elon Musk and Gwynne Shotwell is one of the strongest industrial execution teams in the world over the past two decades. Falcon reuse, Dragon commercial crew, Starlink commercialization, and the governmentization of Starshield all prove that past capital allocation was not wishful thinking, but the realization of many high-return projects. On this dimension, management's historical track record is close to top-tier.

But from the standpoint of a "long-term minority shareholder," this conclusion must be discounted. The IPO documents Reuters has seen show that after listing, SpaceX plans to adopt a dual-class share structure, with Musk and a small group of insiders holding super-voting rights; the documents also include arrangements that limit shareholders' ability to influence board re-election and pursue legal recourse. This means what you buy is likely a ticket to share in economic outcomes, not an ownership certificate that lets you effectively influence capital allocation. For a Buffett-style investor, this is not ideal governance.

The bigger issue is xAI consolidation. Reuters reports that in 2025 the combined entity's total capital expenditure reached $20.74 billion, of which 61% belonged to the AI business; the AI segment posted an operating loss of $6.4 billion; and Starlink's operating profit of $4.42 billion largely played the role of "transfusing blood into AI." In other words, the once-cleaner "spaceflight plus satellite communications" investment logic has been actively reshaped by management into the logic of "spaceflight/satellite cash flow supporting a long-horizon AI narrative." This may be ambitious from an entrepreneur's perspective, but from a minority shareholder's perspective it has already drifted away from the conservative definition of "rational capital allocation."

So my composite score for management is: operating capability 5/5, capital-allocation discipline 2/5, minority-shareholder friendliness 1/5, composite 2.5/5. This is not to say management is not smart, nor that they will not create enormous value; it is to say they look more like people who "allocate capital for mission and scale" than people who "allocate capital for per-share intrinsic value growth." Those two are not always aligned.

Financial Quality and Owner Earnings

Verifiable Financial Quality

The conclusion first: the financial quality of SpaceX's core business is markedly better than that of the vast majority of listed space companies; but the post-merger public financial profile is clearly worse than that of a typical high-quality value stock. That is because core SpaceX has already proven revenue scale, profitability, and industrial dominance, while xAI consolidation has substantially raised losses, capital spending, and governance complexity.

The table below summarizes the key financial and operating metrics that can be relatively verified today. To be clear: the 2021-2024 revenue figures come mainly from media/public documents cited by Reuters, not from the formal figures in publicly downloadable audited statements; for 2025, one must distinguish the standalone SpaceX figures from the consolidated xAI figures.

The data above point to several very important facts. First, revenue growth is extremely fast: on the Reuters-aggregated basis, revenue rose from about $2.3 billion in 2021 to about $13.1 billion in 2024, and standalone SpaceX had reached the $15-16 billion range in 2025. Second, core profitability is strong: standalone SpaceX 2025 EBITDA was about $8 billion; Starlink alone contributed $4.42 billion in operating profit. Third, post-merger cash-flow quality has clearly deteriorated: 2025 total capital expenditure of $20.74 billion was even higher than the $18.67 billion of revenue that year. Fourth, the balance sheet is not fragile in the short term: consolidated year-end cash was $24.8 billion, meaning it will not run into trouble from a liquidity squeeze in the near term, but that is not a synonym for "high-quality cash cow"; it merely reflects "very strong access to financing."

Many of the traditional metrics you asked for, I must honestly write down as unknown. These include: operating cash flow over the past 5-10 years, free cash flow, free-cash-flow conversion, ROE, ROIC, ROA, net debt / EBITDA, interest coverage, maintenance capital expenditure, the full historical trend of receivables/payables/inventory, the complete picture of share dilution, and so on. As of today, the public still cannot obtain the complete audited detail. For a value investor, this in itself is an important negative signal: when you are valuing a giant, complex enterprise, the level of information transparency still falls short.

On the question of "is the profit real cash profit, or accounting profit," my judgment is: Core SpaceX's profit is closer to real operating profit; the post-merger entity's profit quality is markedly lower. The evidence is that Starlink's operating profit is already sufficient to cover the spaceflight segment's losses, showing that the core satellite-internet business has genuine commercial viability; but at the same time, post-merger total capital expenditure far exceeds revenue, and the AI segment posts enormous losses, showing that the free cash flow distributable to shareholders is still being heavily compressed. As a result, this company today looks more like an enterprise that "owns a very strong money-making machine but continues to plow the money it earns into uncertain projects on a large scale," rather than one that "already steadily generates high free cash flow."

On financial fraud or aggressive accounting, I have not seen direct evidence. But I also cannot offer a high-confidence conclusion that "there are no problems," because the complete audited disclosure has not yet been released. The more precise way to put it is: there are no obvious fraud red flags, but the lack of disclosure transparency is itself a diluted version of a red flag.

Owner Earnings Analysis

I must first note the limits of the method. Buffett's ideal Owner Earnings formula is: Net income + non-cash charges - maintenance capital expenditure ± changes in working capital. The problem is that SpaceX currently does not disclose historical statements detailed enough for us to precisely break out these items; the estimates below therefore fall under 【Assumption】 + 【Inference】, not precise accounting fact.

I will use a more conservative framework, one that better fits an owner's mindset, to assess the true earning power of "core SpaceX": 【Fact】 Standalone SpaceX 2025 EBITDA was about $8 billion. 【Fact】 The combined entity's total capital expenditure was $20.74 billion, of which the AI segment accounted for 61%, implying non-AI capital expenditure of roughly $8 billion. 【Fact】 Cumulative investment in Starship has exceeded $15 billion, and 2025 space-segment R&D was $3 billion. Based on these facts, I believe a sizable share of current non-AI capital expenditure is growth investment, but it is certainly not all "negligible expansionary capital expenditure," because Starlink satellite replacement, ground stations, terminals, launch infrastructure, and continued R&D include a good deal of money that must be spent continuously to maintain competitive position.

My conservative estimate is as follows: 【Assumption】 Maintenance capital expenditure is about half of non-AI capital expenditure, roughly $4 billion. 【Assumption】 Cash taxes, interest, and the working-capital drag total $1 to 1.5 billion. 【Inference】 Core SpaceX's current Owner Earnings is therefore only about $2 to 4 billion, and the conservative point estimate I use for valuation is $2.5 billion. This means: even though the core business is very good, it remains a distance away from a mature cash cow that "can freely distribute most of its profit to shareholders." It looks more like a high-quality engine, but for now that engine has not turned its power into large amounts of discretionary cash in shareholders' pockets; instead it keeps pouring that power into Starship and longer-horizon projects.

Viewed this way, SpaceX's problem becomes very clear: it is not that it cannot make money; it is that it still depends heavily on reinvestment, and much of that reinvestment has not yet proven its return. That differs from the typical Buffett-style enterprise with "high ROIC, light capital expansion, and strong cash-distribution capability." For exactly this reason, I will not use "a very large future TAM" as a substitute for "high enough owner earnings today."

Intrinsic Value and Margin of Safety

Intrinsic Value Estimate

I will value the company three ways, but the conclusion first: no matter which method I use, the currently verifiable valuation range shows no adequate margin of safety.

Method one: discounted Owner Earnings. This is the method I weight most heavily, because it is closest to the logic of "buying an entire business." Based on the conservative Owner Earnings estimate above, I lay out three scenarios:

• Conservative scenario: starting Owner Earnings of about $2.5 to 3 billion; 10-year annualized growth of 15%; discount rate of 12%; terminal growth of 3%. The corresponding intrinsic value is roughly $80 billion to $150 billion.

• Neutral scenario: starting Owner Earnings of about $4 to 5 billion; 10-year annualized growth of 18%; discount rate of 10%; terminal growth of 4%. The corresponding intrinsic value is roughly $180 billion to $320 billion.

• Optimistic scenario: starting Owner Earnings of about $6 to 7 billion; 10-year annualized growth of 22%; discount rate of 8%; terminal growth of 5%, assuming Starship economics, direct-to-cell, and a larger-scale Starlink constellation all materialize smoothly. The corresponding intrinsic value is roughly $500 billion to $800 billion.

The core inputs to these scenarios come from facts such as standalone SpaceX's roughly $8 billion EBITDA, Starlink's $4.42 billion operating profit, total capital expenditure of $20.74 billion with the AI segment at 61%, and Starship's large, continued investment; but the valuation itself is my judgment, not market consensus.

Method two: relative valuation. This method is not friendly to SpaceX, because peers diverge too widely and most of them are themselves expensive. Among listed comparables, Iridium's 2025 revenue was $871.7 million, OEBITDA $495.3 million, and net debt about $1.7 billion; on current market value, that works out to roughly 7.0x EV/Sales and about 12.4x EV/OEBITDA. Rocket Lab's 2025 revenue was $601.8 million, with year-end cash and marketable securities of about $1.099 billion and book debt of about $154 million; on current market value, EV/Sales is roughly 124x, yet it still loses money. AST SpaceMobile's 2025 revenue was $70.9 million, with year-end cash of about $2.336 billion and debt of about $2.22 billion; on current market value, EV/Sales runs as high as 341x, and it too is not yet steadily profitable. In other words, the public-market "space stocks" are themselves riddled with narrative premium, and the fact that peers are also expensive does not prove SpaceX is cheap.

Now look at SpaceX itself. Per Reuters, standalone SpaceX 2025 revenue was about $15-16 billion and EBITDA about $8 billion. At the $1 trillion pre-merger valuation as of February 2026, that is about 62.5-66.7x sales and 125x EBITDA; at Scottish Mortgage's carrying value of $1.25 trillion, the sales multiple is higher still; at the Reuters-reported $1.54 trillion secondary-market valuation or the $1.75 trillion IPO target, the pricing logic moves closer to that of a "mega-cap AI platform company" rather than a "mature but high-growth spaceflight/communications infrastructure company." Even after Reuters adopts very aggressive, undervaluation-leaning assumptions, assuming cash flow and revenue double in 2026, it still calculates that $1.75 trillion corresponds to about 56x revenue and 109x EBITDA, and states plainly that "multiples are stretched."

Method three: asset or liquidation value. For a growth company, this method usually gives you a floor, not a ceiling. The combined entity's year-end 2025 data that Reuters has seen show SpaceX with $92 billion in total assets and $50.8 billion in total liabilities, for a book net worth of about $41.2 billion. If the market prices it at $1.25 trillion, that is about 30x book net worth; at $1.75 trillion, it approaches 42x book net worth. This tells us: the asset method at most tells you "it is not an empty shell," but it absolutely cannot provide adequate protection for the current price.

Combining the three methods, the conclusion I reach is: Conservative intrinsic value range: $100 billion to $200 billion. Fair intrinsic value range: $200 billion to $350 billion. Optimistic intrinsic value range: $500 billion to $800 billion.

Under this definition, the currently verifiable valuation range of $1.25 to $1.54 trillion, and even the rumored $1.75 trillion, all sit well above my fair-value range. Put differently, the current price looks like paying in advance for the full-marks script of "Starship succeeds + Starlink keeps booming + xAI is ultimately monetized + minority shareholders are not harmed." For a value investor, that is not a margin of safety; it is the opposite of one.

So the price judgment I give is: Ideal buy price range: $150 billion to $250 billion. Acceptable long-term holding price range: $250 billion to $400 billion. Clearly overvalued range: above $600 billion; especially above $1 trillion. This is not to say the company is only worth this much; it is to say that when you cannot verify the next decade's cash flow with high confidence, Buffett-style discipline does not permit you to pay far above the "operating facts already realized."

Margin of Safety

The heart of a margin of safety is not "is the company great enough," but "even if the future is not so great, will you avoid losing too badly." SpaceX's single biggest problem right now is exactly this: a good company, but most likely a bad price. As long as your premise at the time of purchase is "the future is sure to be great," you simply do not have a margin of safety.

At the current valuation, there are at least three most-fragile assumptions. First, Starship must not only fly, but also be economically viable; otherwise the large capital investment will not convert into high returns. Second, Starlink must keep expanding while maintaining relatively high profitability, because it has in effect already taken on the role of transfusing blood into both the spaceflight segment and the AI segment. Third, the xAI-related investment must ultimately create returns, or minority shareholders will keep bearing the valuation compression caused by large-scale capital consumption. If any one of these three clearly fails to hold, the return at the current price could weaken substantially.

If growth comes in below expectations, margins decline, or the valuation multiple reverts to something closer to an infrastructure/communications company, permanent capital loss will appear. Taking the currently verifiable valuation of about $1.25 to $1.54 trillion as the reference point, if over the next decade the market is ultimately willing to assign this company only a valuation range closer to "a good business but not a myth," such as $300 billion to $500 billion, then even if the enterprise itself keeps growing, shareholders could still suffer a 50% to 80% drawdown in market value. This is not "short-term volatility"; it is the long-term compounding damage caused by paying too high an entry price. This conclusion is my 【Inference】, but it rests directly on the known facts of today's high multiples and heavy capital spending.

My clear judgment is: the current margin of safety is inadequate, and very much so. If you hold to Buffett-style discipline, the most reasonable action is not "getting on board before a great company lists," but acknowledging that there is no sufficiently cheap price in front of you. Waiting is itself an investment action.

Risks, the Bear Case, and Comparisons

Risks and the Strongest Bear Case

I rank SpaceX's important risks in order of "likelihood of causing permanent capital loss," roughly as follows.

First is valuation risk. This is the first risk, not the sixth. Because the company is great enough, it becomes all the easier for investors to ignore the price. The market is already pricing a "spaceflight/communications + AI optionality" hybrid the way it would price a mega-cap AI platform, and that valuation still embeds a great deal of capability that has not yet materialized.

Second is technology and execution risk, pointing chiefly at Starship. Cumulative investment in Starship exceeds $15 billion, and 2025 space-segment R&D was $3 billion, but the offering materials Reuters cites are quite direct: to reach the "thousands of launches a year" Musk talks about, SpaceX still faces a series of not-fully-solved problems in thermal protection, ground facilities, water resources, fuel supply, and orbital refueling, and the company itself admits these strategic initiatives "may not be achieved on schedule, or may not be achievable at all."

Third is regulatory and government-dependence risk. SpaceX's advantage and its risk partly come from the same source: it is deeply tied to the FAA, FCC, NASA, Space Force, NRO, and similar systems. FAA approvals, FCC spectrum, large government orders, and national security programs give it a moat, and also leave it highly exposed to shifts in policy, budgets, geopolitics, and the founder's relationship with the government.

Fourth is governance and capital-allocation risk. The dual-class arrangement means that even as a shareholder you can hardly constrain management; and xAI consolidation has already proven that management can reshape a once-more-comprehensible investment target into one with heavier capital spending, more complex accounting, and a grander narrative. For minority shareholders, this is not a minor issue; it is precisely the reason the discount rate in the pricing model itself must be raised.

Fifth is competition risk. Although SpaceX still leads, Amazon Leo, Eutelsat OneWeb, AST SpaceMobile, Blue Origin, and Rocket Lab are all closing in on or carving into the profit pool at different points. Especially in low-orbit broadband and direct-to-cell, the future may not be "winner takes all," but more likely "the strong stay ahead while margins are ground thin by competition."

The strongest bear case can be condensed into a single sentence: You think you are buying a great company that has already proven its business model, but you are more likely paying an extremely high price to pre-fund an entire future that has not been adequately audited and that includes a long-horizon AI narrative.

Investors who are bearish on SpaceX are most likely to see this: it is of course an outstanding enterprise, but an outstanding enterprise does not automatically equal an outstanding stock; Starlink's success does not necessarily mean Starship's economics will hold; a strong record of past capital allocation does not necessarily mean capital allocation after consolidating xAI is still friendly to minority shareholders; the current valuation has already written too many favorable outcomes into the price.

The facts that would make me admit I am wrong and turn more optimistic include: Starship significantly improving reliable reflight and per-payload economics over the next 2-3 years; Starlink continuing to prove, from a high base, that ARPU, penetration, and government/enterprise expansion advance together; xAI/AI investment beginning to generate verifiable returns rather than continuously consuming Starlink's profit; post-IPO disclosure showing segment cash flow better than current inference, with related-party transactions and governance arrangements more restrained than expected. Conversely, if Starship cannot reach key milestones for a long time, Starlink growth slows markedly, AI capital spending outpaces revenue growth for an extended period, or minority shareholders are continuously diluted/marginalized, then I would more firmly conclude that "this company's greatness is not enough to support the current investment."

Comparison With Other Opportunities

If your goal is to allocate capital toward "space/satellites," Rocket Lab is the more direct listed pure-play substitute, but its 2025 revenue was only $600 million, it still recorded a net loss, and its current EV/Sales of about 124x is not cheap. By comparison, SpaceX's business quality and scale are of course better, but that does not mean buying at a far higher market value is a better deal.

If what you want is "a satellite-communications company that has already proven it can make money," Iridium is actually closer to a traditional value framework. Its 2025 revenue was $871.7 million, OEBITDA $495.3 million, and net income $114.4 million, and it is buying back stock; its current P/E is about 42x and EV/OEBITDA about 12.4x. Iridium's growth is far below SpaceX's, but its cash flow, governance, and valuation are all closer to explainable boundaries.

If your core interest is actually "satellite internet + large tech-platform optionality," then Amazon is also a substitute that cannot be ignored. Official materials show Amazon Leo is deploying thousands of satellites and plans to expand service in 2026, but what you buy with Amazon is a diversified giant with cash flows from AWS, e-commerce, advertising, and cloud infrastructure, not a single spaceflight theme. Amazon's current P/E is about 31.6x. This does not mean it will necessarily return more than SpaceX, but it does mean that if you merely want to participate in "the future of space infrastructure," you do not have to pay the hottest price for the hottest name.

Against a broad-based index, the S&P 500 is described by S&P Global as covering about 80% of investable U.S. market value; against the risk-free rate, FRED shows the U.S. 10-year Treasury yield recently around 4.47%. At the current price, SpaceX, to me, does not offer a risk premium clearly superior to a broad-based index and medium-to-high-grade U.S. Treasuries: you are taking on single-company risk, governance risk, technology risk, and valuation risk, yet in my neutral scenario the expected return is not meaningfully higher than these alternatives.

So if I could hold only 5 assets, SpaceX at the current price does not qualify for the portfolio. Not because it is not excellent, but because capital is scarce, and an excellent-but-very-expensive asset should not automatically crowd out an asset that is "good enough and more reasonably priced."

Investment Checklist and Final Verdict

Investment Checklist

The table below is the "pass / fail / uncertain" checklist you requested:

Final Investment Conclusion

【Final Rating】 Watch

【One-Sentence Investment Thesis】 SpaceX is a truly great core enterprise, but within the current verifiable/rumored valuation range of $1.25 to $1.75 trillion, it looks more like a high-priced speculation on future optionality than a value investment with a margin of safety.

【Core Bull Case】

• Core business quality is extremely high: launch, orbital logistics, satellite internet, and government security networks have already proven commercial viability.

• The moat is real and deep: reuse, manufacturing scale, internal payload demand, government credentials, and execution culture together form a system that is hard to replicate.

• Starlink has become a high-quality profit engine, with 2025 operating profit of $4.42 billion.

• Government and defense contracts are deep, supporting demand stability and industry position.

• If Starship truly achieves large-scale, low-cost reflight, the enterprise's ceiling could be far above today's conservative estimate.

【Core Bear Case】

• The current valuation substantially front-runs the future and lacks a margin of safety.

• The listing entity has already consolidated xAI, clearly rewriting the financial quality and investment logic.

• Total capital expenditure of $20.74 billion exceeds that year's revenue, showing that free cash flow is not friendly to shareholders.

• The dual-class structure and restrictions on shareholder rights weaken minority-shareholder protection.

• The complete audited prospectus has not yet been released, and many key financial metrics still cannot be verified with high confidence.

【Key Assumptions】

• Starlink can still sustain high growth and good margins from a high base.

• Starship truly crosses the threshold from "technically able to fly" to "economically replicable" over the next several years.

• The xAI/AI infrastructure investment ultimately produces verifiable returns rather than continuously consuming cash.

• Government relations, regulatory approvals, and key contracts do not deteriorate structurally.

• Post-listing disclosure transparency improves, and capital allocation does not continue to deteriorate for minority shareholders.

【Fair Buy Price】 An enterprise valuation range of $150 billion to $250 billion better fits a conservative Owner Earnings and margin-of-safety requirement; $250 billion to $400 billion can be understood as a "great company but not cheap" holding range; above $600 billion I would define as clearly overvalued; above $1 trillion is closer to paying for a legend.

【Target Holding Period】 If a suitable price truly appears in the future, at least 10 years. But the premise is that the entry price itself must allow you to earn a reasonable return even when the script is not full marks.

【Expected Annualized Return】 This assumes you are buying near the current verifiable valuation range.

• Conservative scenario: -12% to -8%. Corresponding to slow realization of Starship economics, slowing Starlink growth, and valuation contraction.

• Neutral scenario: -1% to +2%. Corresponding to the company staying excellent, but the market ultimately pricing it as "an excellent infrastructure company" rather than "a mythic platform."

• Optimistic scenario: +8% to +11%. Corresponding to Starship, Starlink, and AI optionality all broadly materializing. For a value investor, this is the crux: the neutral scenario is not attractive enough, and the bad scenario's loss is too large.

【Maximum Loss Risk】 70% to 85%. In the worst case, the company is still a good company, but the market re-rates it from "the peak of the narrative" back to "an excellent but capital-intensive infrastructure enterprise," and you bought at an extremely high valuation, which produces a very typical permanent capital loss.

【Tracking Metrics】

• Starlink active customers, ARPU, and segment margin.

• Starlink terminal cost and satellite replacement cadence.

• Starship reliability, reflight interval, per-launch cost, and orbital-refueling progress.

• The split between non-AI core capital expenditure and maintenance capital expenditure.

• xAI segment revenue, loss, and capital-expenditure growth rate.

• The cadence of government contract renewals and new awards, especially NSSL, NRO, and NASA.

• Key FCC/FAA approvals, especially direct-to-cell spectrum and Starship launch cadence.

• Complete segment financial disclosure after the IPO.

• Changes in share dilution, related-party transactions, incentive structure, and the dual-class arrangement.

• Whether management continues to shift Starlink cash flow into high-uncertainty projects.

【Signals That Trigger Reassessment】

• The public prospectus shows Owner Earnings clearly above the current inference.

• Starship achieves a substantive breakthrough at the economics level.

• Starlink's growth and margins both substantially beat expectations.

• xAI begins to form verifiable returns rather than continuing to burn cash.

• Management improves minority-shareholder governance arrangements.

• Or the reverse: regulatory obstruction, contract delays, continued runaway capital spending, slowing growth, and a lack of disclosure transparency.

【Final Recommendation】 Viewed calmly, with restraint, and with a long-term orientation, SpaceX is worth admiring, but it is not worth chasing right now. If you hold to a Buffett-style framework, the most reasonable conclusion is not "it would be a pity to miss it," but "great enterprises are many; entry points with a margin of safety are few." At today's price and this level of disclosure, I would put SpaceX on the high-priority watch list, not the buy list.

This report is based on public information and does not constitute investment advice. Markets carry risk; invest with caution.