SpaceX: A Deep-Dive Research Report

The first true turn of fate came in 2008. That year NASA awarded International Space Station commercial resupply contracts to SpaceX and Orbital separately, with SpaceX winning $1.6 billion across 12 missions. The significance lay not only in the dollar figure but in pushing SpaceX from an engineering startup "still proving it could survive" into a delivery contractor backed by national credit. In hindsight this step was crucial: without the early NASA contract, SpaceX could hardly have lasted through a long R&D cycle; and once it had NASA, its credibility in pursuing later defense and intelligence orders was no longer zero.

The second phase was the decisive contest of industrialization and reusability. Falcon 9's commercial meaning is not merely "it can launch" but that it turned commercial space from a low-frequency, expensive, one-off project into a high-frequency, relatively standardized delivery system that looks more and more like an industrial product. SpaceX's official materials show Falcon 9's current public list price is $74 million, the rideshare program can do 50 kg for $350,000 at roughly $7,000 per kg of incremental mass, and the official launch page already shows a single first-stage booster flying to its 30th mission. This is not marketing spin but industrial discipline: when a heavy asset that once had to be scrapped after a single use becomes a "fleet" reused dozens of times, the industry's cost curve gets rewritten.

The third phase was growing from a "launch company" into a "platform company." In 2020, NASA's Demo-2 marked SpaceX's historic milestone in commercial crewed spaceflight; in 2021, NASA awarded SpaceX the $2.89 billion lunar landing system contract, making Starship no longer just a Mars fantasy but part of America's deep-space program; in 2023, Musk publicly stated that Starlink had reached cash-flow breakeven; and by 2024, SpaceX completed 134 Falcon launches, becoming the world's most active launch operator. The throughline matters: the company was not going deeper on a single track but repeatedly proving one capability across launch, crew, defense, and broadband — turning extraordinarily complex space systems into sustainably delivered products.

The fourth phase is the fundamental change of 2026: SpaceX is no longer just SpaceX. In February, SpaceX merged with xAI at a combined valuation of $1.25 trillion; the IPO summary Reuters saw in April already showed the AI business with an operating loss of roughly $6.36 billion in 2025, with capex devouring a large share of the money Starlink earned; on May 20 the S-1 went public, with the company's target valuation lifted to $1.75 trillion and an equity structure designed so Musk still holds 85.1% of the voting power after the IPO. In other words, this listing is not "a SpaceX stock you can finally buy" but a capital event that packages rockets, satellite internet, AI infrastructure, and Musk's control and hands the whole bundle to the public market.

The table below compresses the historical narrative into the handful of nodes that truly changed the company's fate:

Sources: company website, NASA, Reuters.

If you fold the "share price and valuation history" into this timeline, the logic grows clearer: in 2023 it was still "the scarcest private space stock," in 2025 it became "the private-market king whose valuation soared rapidly after Starlink proved the model," and in 2026 it was packaged as "a super tech-infrastructure asset that can enter public markets." This is not valuation mean reversion in the traditional sense but capital markets changing the very definition of the company's category. Precisely for that reason, SpaceX's valuation does not rise smoothly along revenue and earnings but jumps along the boundary of its narrative.

Business Model, Historical Financial Review, and Moats

SpaceX's commercial machine today is no longer as simple as "selling rocket launch services." Per the 2026 S-1, 2025 consolidated revenue was about $18.67 billion, made up of three segments: Connectivity at roughly $11.4 billion, Space at roughly $4.1 billion, and AI at roughly $3.2 billion; in Q1 2026 the three segments brought in roughly $3.257 billion, $619 million, and $818 million respectively. This means the biggest earner is not the most dramatic Starship but the Starlink that most resembles a telecom subscription business. Reuters' reporting on the prospectus summary also shows the connectivity business that houses Starlink earned about $4.423 billion in operating profit in 2025, nearly covering the Space segment's loss, and continued to contribute $1.19 billion in operating profit in Q1 2026; by contrast, the AI business posted an operating loss of about $6.355 billion in 2025 and lost another $2.47 billion in Q1 2026.

The economics of the three segments are entirely different. Connectivity is deferred investment with recursive monetization: satellites go up via rockets up front, then subscriptions and service revenue come back through home broadband, enterprise lines, aviation, maritime, government, and direct-to-cell. The official public map shows Starlink already covers 150+ countries/markets; the prospectus summary shows about 10.3 million Starlink subscribers at the end of March 2026, coverage in 164 countries and regions, and roughly 9,600 to 10,000 satellites in orbit. This business most resembles a traditional platform: the more units installed, the broader the coverage, the lower the unit cost, and the better the marginal profit. And once that same network starts supporting aviation, maritime, government, and direct-to-cell, pricing power and ARPU structure shift from single-consumer broadband toward higher-value scenarios.

The Space business is more of a project-plus-platform hybrid. It includes Falcon 9/Falcon Heavy launches, Dragon cargo and crew, national-security missions, Starshield, and the still heavily-invested Starship. Margins here will not improve linearly the way SaaS does, because it must sustain high investment in manufacturing, testing, launch sites, and R&D facilities over the long term; but once launch cadence and reuse rates keep rising, it generates very strong operating leverage. Official materials show Falcon 9's maximum payload to LEO is 22,000 kg at a list price of $74 million; and BryceTech's data shows that of the 11,080 small satellites launched globally from 2020 to 2024, 69% were launched by Falcon 9. This means the real moat of the Space segment is not "each launch is expensive" but that others run projects while it runs liners and fleets.

The AI business is the hardest to value and the most dangerous part today. Per Reuters and the S-1 summary, the company's total capex in 2025 was about $20.74 billion, of which 61% belonged to AI; Q1 2026 capex already reached $10.1 billion, of which 76% was still AI. This makes the post-IPO SpaceX no longer a company that "supports expansion with profits" but one that "transfuses the loss-making segment with the profitable segment, and keeps refueling from the capital markets." Bulls will call this an active capital allocation that converts Starlink's cash flow into a larger arena; bears will say investors are buying not pure SpaceX but a super holding structure whose valuation anchor is being devoured by AI. Both views rest on facts.

From a balance-sheet view, the thing to watch is not "no business on the books" but whether enormous debt and capex will squeeze free cash flow negative and keep it there too long. The prospectus summary shows that as of March 31, 2026, the company's total principal interest-bearing debt was about $29.132 billion. Reuters' April exclusive separately mentioned that a $20 billion bridge loan had brought early-March debt down to about $20.07 billion. The two figures do not fully match, most likely because of differing observation points and accounting bases; but this does not fundamentally change the research conclusion: the official S-1 quarter-end figure should serve as the baseline, and SpaceX's leverage is not light. This is not the debt problem of a distressed company but the classic refinancing problem of a high-momentum, high-capex, high-narrative company.

I see four moats that genuinely hold. The first is the cost advantage and learning curve from reusability. Not every rocket company can fly a first-stage booster to 30 flights, and not every company can stack launch density to today's level. The second is vertical integration: rockets, satellites, ground stations, terminals, launch sites, and government delivery all sit inside its own system, which is efficient and fast to feed back. The third is network scale and service density: 10-million-plus subscribers and 150+ countries of coverage turn Starlink from a "satellite broadband product" into a "global connectivity network." The fourth is government and national-security trust: NASA resupply, commercial crew, the lunar landing system, the Space Force's national-security launch contracts, and Starshield were not won with marketing copy. At this point SpaceX's competitors are no longer only other space startups but increasingly come from traditional telecom, defense, and cloud infrastructure.

Compared with business quality, governance is its most obvious weakness. Reuters states plainly that after the IPO Musk will retain 85.1% of the voting power; Class B carries 10 votes per share and Class A 1 vote per share; the company also sets a series of provisions to limit common-shareholder rights, including mandatory arbitration, restricted venue for litigation, and even prospectus design that "protects Musk from being fired by anyone other than Musk." Judged purely on execution, Musk is SpaceX's biggest asset; judged on public-company governance, he is a source of discount. This structure — "the person is the moat and also the governance risk" — is rarely taken to such an extreme by any other U.S.-listed company.

The table below compresses the most critical operating and financial handles into one view:

Sources: SpaceX S-1 summary and Reuters.

Industry, Cycle, and Peer Comparison

If a single competitive-landscape scenario must be chosen first, SpaceX belongs to "no direct comparable, but several partial reference points." The reason is simple: it spans launch, crew, satellite internet, national-security networks, and AI infrastructure all at once, while any single listed company can cover only one corner. Rocket Lab is the closest along the "commercial launch plus space manufacturing" line, but it is far smaller; Viasat, Iridium, and Eutelsat/OneWeb are more like indirect comparables in satellite connectivity, but with different technology paths and end markets; AST SpaceMobile is more of a frontier challenger in direct-to-cell; and Amazon is the truly heavyweight alternative behind Kuiper, but it is not a pure space asset. For exactly this reason, SpaceX's high premium is not only about high quality but also about the lack of a comparable.

Consider its industry position first. BryceTech's 2024 global launch statistics show 259 orbital launches that year deploying nearly 2,900 spacecraft, of which roughly 60% were performed by U.S. launch providers and about 70% were commercial launches. On the small-satellite dimension, BryceTech gives two even more striking figures: since 2019 the U.S. has launched the majority of small satellites, with the non-U.S. share falling below 15%; and from 2020 to 2024 Falcon 9 alone carried 69% of small-satellite launches. This explains why SpaceX's competition is not "many identical players in the industry" but "it first turned the industry into an oligopoly, then left others to fight over the edges."

On cyclical character, SpaceX is not a traditional macro-cyclical stock, but it is by no means non-cyclical either. It is exposed to four cycles at once: the policy/budget cycle, the capex cycle, the technology iteration cycle, and the regulatory cycle. NASA and Space Force orders depend on the federal budget; Starship and AI data centers depend on long-cycle capital investment; Starlink's terminal penetration and direct-to-cell sit in the early stage of technology iteration; and launch licensing, spectrum licensing, orbital-debris rules, and national-security review form the regulatory boundary. In other words, SpaceX's volatility is not the old-style industrial "boom expansion and recession" swing but a composite volatility driven jointly by policy, licensing, test success rates, and capital-market risk appetite.

Cross-sectionally, what each peer "grew up to be" differs greatly. Rocket Lab is the listed sample most like a "second-tier SpaceX": it had $602 million in revenue in 2025 and $200.3 million in Q1 2026, growing fast, with a backlog of $2.2 billion, but its core is still small-to-mid launch and space systems, lacking a global subscription asset like Starlink. It is more a high-growth space manufacturing and mission-integration company than a global connectivity network. Viasat is another shape: fiscal 2025 revenue of $4.52 billion with communications-service revenue of $3.298 billion, but its model leans toward geostationary satellites and government/aviation communications — asset-heavy, debt-heavy, slow-growth — so the capital markets give it a multiple far below SpaceX's. Iridium's profile is different again: $871.7 million in 2025 revenue, with solid operating milestones, steady government business, and a reliable network, but it is essentially a narrowband and mission-critical connectivity player whose imaginative ceiling is far below Starlink's, though it wins on a clear model and controllable capex. AST SpaceMobile represents the "concept extremely strong, delivery still distant" direct-to-cell challenger: only $70.9 million in 2025 revenue, yet already handed a stunningly high valuation by the market, showing how aggressively capital markets now price the long-term potential of "space-based communications replacing terrestrial."

Putting current market caps and revenue side by side illustrates SpaceX's valuation tier more vividly:

Note: market caps in the table are approximate values around 2026-05-23; SpaceX is the proposed IPO target valuation; multiples are simplified market cap/revenue, not strict EV/Sales.

What is most worth reading in this table is not "who is more expensive" but why the market produces these differences. The high multiples on Rocket Lab and AST reflect that, while "the next-stage winner in space is not yet clear," capital is willing to pay an exaggerated premium for the incremental story; the low multiples on Viasat and Iridium show capital markets pricing mature satellite communications more like defensive stocks. SpaceX's problem is that it wants both benefits at once: it wants the validation of a mature connectivity network and the long-dated option of the most cutting-edge AI/Starship. If execution succeeds, it could of course sustain a premium for the long term; if execution falls short, the de-rating will also be sharper than for a mature satellite-communications firm.

So SpaceX's real niche is not "simply the leader" but a pump siphoning across cross-industry profit pools. Upstream it grabs share from traditional launch providers, midstream it grabs the future growth of satellite-communications operators, and downstream it pushes through direct-to-cell and AI infrastructure against the boundaries of telecom and cloud computing. Whose profit pool does it most directly take? Near term it is commercial launch and satellite broadband; medium term it is aviation/maritime/government connectivity; and only long term is it terrestrial communications and compute infrastructure. The real worry is not who will immediately take its main battlefield, but whether, in attacking too many fronts at once, it will dilute the most profitable Starlink cash flow to the point where it cannot cover the capital needs of the rest of the business.

Current Fundamentals and the Bull-Bear Debate

Start with the most important reality: SpaceX today is not "performing poorly" but rather "the money the good businesses make is not yet enough to cover what the bad and new businesses spend." In Q1 2026 the company had $4.69 billion in revenue but a $1.94 billion operating loss; within that, the connectivity business that houses Starlink earned $1.19 billion in operating profit, while the AI business lost $2.47 billion in the single quarter, with quarterly capex as high as $10.1 billion. The most misleading thing about this kind of statement is that, without segmenting, you would see a "growing but loss-making" big-tech company; once you segment, you find a connectivity network inside that is already quite profitable.

Consider the operating proxies next. By the end of March 2026, SpaceX had completed about 620 orbital launches with a mission success rate above 99%; Falcon launched 134 times in 2024, a record, and Falcon 9's single-year mission count reached 165 in 2025. On May 22, the upgraded Starship V3 completed its 12th test flight, deploying 20 mock Starlink satellites and 2 real diagnostic satellites, and completed a controlled splashdown. Although the booster stage still wasn't perfectly recovered and there were engine failures, this test flight mattered greatly for pre-IPO market sentiment, because it at least proved Starship has not stalled at the "promising the moon year after year" stage. Conversely, in February Falcon 9 was briefly grounded over a second-stage issue, reminding investors that even the most mature rocket system remains a business highly sensitive to regulation and safety.

The theme the market trades today is therefore not a single "space boom" but five narratives stacked together: the largest IPO, the Musk halo, Starlink's real profitability, Starship's cost revolution, and the ultra-long-dated option on AI infrastructure. In any ordinary company, the simultaneous appearance of "poor governance, exploding capex, extremely high valuation, and an enormous long-term story" would usually be a risk combination; but on SpaceX, these factors are precisely why capital is willing to chase it. One thing the analysis must stay alert to: when capital markets price SpaceX, they may not clearly distinguish what has already been realized in the fundamentals from what is only future imagination.

The bulls' strongest evidence chain has three links. First, Starlink has moved from "telling a story" to a "cash-flow machine": it reached cash-flow breakeven in 2023, the connectivity business earned $4.423 billion in operating profit in 2025, and subscribers exceeded 10 million by the end of March 2026. Second, the launch moat has been industrially proven: a single Falcon 9 first-stage booster flew to 30 flights, Falcon 9 carried 69% of global small-satellite launches from 2020 to 2024, and it was allotted 28 missions worth $5.9 billion under the hardest national-security program, NSSL Phase 3 Lane 2. Third, the optionality is enormous and scarce: direct-to-cell received FCC approval, with another 7,500 second-generation Starlink expansion licenses; and after EchoStar's 65 MHz spectrum deal was approved, Starlink holds even more cards in hybrid cellular/satellite networks.

The bears' strongest evidence chain also has three links. First, the current valuation anchor is extremely weak: $1.75 trillion against $18.67 billion of 2025 revenue corresponds to roughly 94x P/S; judged only on already-proven current cash flow, this valuation can scarcely be supported by Starlink and launch alone. Second, the AI consolidation has markedly worsened earnings quality and comparability: AI's 2025 operating loss was $6.355 billion, with another $2.47 billion lost in Q1 2026, and capex devoured most of the internal cash flow. Third, the governance provisions give common shareholders almost no real check: 85.1% voting power, dual-class shares, arbitration and litigation-restriction clauses, and Musk's extreme control — these can be papered over by the halo in good times but will amplify the discount in bad times.

Valuation Analysis, Risks, Catalysts, and Tracking Metrics

Because SpaceX is not yet formally listed and the final offering range and fully diluted share count remain incompletely disclosed, giving a precise per-share target price the traditional way would be false precision. This report therefore uses a three-step method: first look at the anchor capital markets have already given, then compare relative to peers, and finally run a low-precision but traceable scenario analysis.

Consider historical valuation first. Over the past three years, market perception of SpaceX shifted gears all the way from "the most expensive private space company" to "potentially one of the most expensive listed tech-infrastructure assets in the world." In 2023 the valuation was still on the order of $150 billion; by the end of 2025 secondary transactions had reached around $400 billion, with talk of an $800 billion equity deal; in February 2026 the merger with xAI carried a combined valuation of $1.25 trillion; and in May 2026 the IPO target valuation rose to $1.75 trillion. This process shows that the change in valuation's center of gravity was driven mainly not by short-term earnings but by an asset-category switch: from space manufacturing, to global connectivity, to an AI/space super-platform.

Consider relative valuation next. Even set against the white-hot new-space/direct-to-cell sector of 2026, a $1.75 trillion target valuation is still very aggressive. It exceeds the multiples of the vast majority of mature satellite-communications companies and the absolute market caps of many high-growth space stocks, and it means the market is granting it not "a premium on the existing business" but a packaged prepayment on Starship commercialization, direct-to-cell penetration, government security-network expansion, and AI platformization. That is why I would rather view it as a super option than an ordinary growth stock.

The scenario table below is not investment advice; it simply unpacks current market expectations for clarity:

Scenario basis: built on the current revenue structure in the prospectus and the capex and segment-profit information Reuters disclosed, with reference to the market pricing of comparable new-space and satellite-communications companies.

The most important conclusion in this table is not the specific numbers but this: to turn the $1.75 trillion valuation into "neutrally fair," SpaceX must deliver at least two things at once — Starlink keeps growing fast, and at least one of Starship/AI moves from story to cash flow. Relying only on "the profit segment already disclosed today," it is hard to grow naturally into that valuation; it must lean on expectations, and what expectations fear most is delay.

On risk, I suggest writing it as trackable variables rather than empty words:

Sources: SpaceX S-1, Reuters, FCC, FAA.

Catalysts work the other way, with four key ones. First, the first quarterly report after listing, which will for the first time let the public market see the durability of Starlink, Space, and AI, not just a prospectus snapshot. Second, the next several Starship test flights, especially milestones closer to commercialization such as recovery, on-orbit refueling, and commercial payload deployment. Third, new direct-to-cell and government-communications contracts, including spectrum consolidation, Starshield, and potential missile-defense and national-security network opportunities. Fourth, whether the capex and debt curves slow — the key to deciding whether it is a growth stock or a high-valuation cash-burn stock.

If I were to build the most practical tracking dashboard, I would watch these eight metrics: Starlink subscribers, connectivity operating margin, AI-segment loss, capex/revenue ratio, interest-bearing debt level, Falcon launch cadence, Starship key test milestones, and new FCC/FAA/NASA/Space Force licenses and contract announcements. Among these, the most important is not revenue growth but whether Starlink can keep making money faster than AI and Starship spend it.

A Cross-Cutting Conclusion

Viewed historically, the capability SpaceX has truly proven is not "telling a good story" but that while others treated space as an engineering project, it built space as an industrial system; and while others treated satellite internet as a test bed, it turned Starlink into a cash machine. Both are extraordinarily hard to replicate. Falcon's reusability and high-frequency launch prove it can drive down unit cost; Starlink's scaled subscriptions prove it can turn one-off capital investment into recurring collections. Most "great hard-tech companies" manage only the first half; SpaceX managed both halves.

Its past success came from five factors stacked together: the tailwind of the era, government orders, management capability, technological edge, and capital-market tolerance. The era's tailwind is the broad trend of low-orbit satellites, national security, and space commercialization; government orders gave it early cash flow, credibility, and a long-term anchor; management capability shows in vertical integration and ultra-fast iteration; the technological edge shows in reusability and constellation operations; and capital-market tolerance gave it the ability to raise at ever-higher valuations again and again. Those five factors have not vanished today, but their weights have shifted: what mattered most in the past was engineering delivery, and what matters most today is capital-allocation discipline. Because when a company simultaneously holds a profitable Starlink, a loss-making Starship, and a heavily-invested AI, what you truly need to judge is no longer just "can it build it" but "will it take the money its best business earns and over-spend it on the most distant story."

Viewed cross-sectionally, SpaceX's real edge over peers is not any single parameter but the systemic combination. Rocket Lab has no Starlink network; Viasat and Iridium have no launch-and-manufacturing chassis like Falcon; AST is still proving commercialization; and Amazon's Kuiper, though richly resourced, is not an independently listed comparable asset. SpaceX's strength is stringing rockets, satellites, terminals, spectrum, government relationships, and manufacturing into one chain, so it can harvest more synergy from the same dollar of capex than others. Its weaknesses are also clear: governance is over-concentrated, the financial accounting grows ever more complex, and the valuation anchor blurs further after AI consolidation. Among these weaknesses, governance is structural, not something that naturally repairs in the short term; financial complexity will not vanish in a single earnings report either.

Is the current valuation rewarding its past success or borrowing against the future? My answer: both, but the borrowing-against-the-future component is larger. If it only rewarded the past, SpaceX would be an extremely high-quality global connectivity and commercial-space leader deserving a significant premium; but when the valuation reaches $1.75 trillion, the market is clearly paying not only for Falcon and Starlink but also prepaying for Starship's full commercialization, large-scale direct-to-cell monetization, orbital AI infrastructure, and even the far more distant space manufacturing and deep-space economy. The catch is that only part of these options has been repeatedly proven.

What the market is most likely to misjudge right now is not "SpaceX won't make it" but packaging assets of different certainty as if they were the same quality. Starlink's profit quality, the Space segment's capital intensity, and the AI segment's loss nature were never the same. Once they are kneaded into one "super tech platform," the market easily overestimates how well the good assets can cover the bad. In the real world, the most common risk is not that the core business suddenly disappears but that the money the core business earns is never enough to fill the holes in the other segments, ultimately dragging the valuation center lower.

The key variable for the coming year is whether the first batch of post-listing earnings can prove that Starlink's growth has not clearly stalled as it scaled, AI's capex has not kept spiraling out of control, and Starship's technical milestones keep advancing. The key variable for the next three years is whether Starship moves from engineering validation to commercial validation, whether direct-to-cell and government security networks form high-value increments, and whether AI moves from a "high-investment lab" into an infrastructure platform that produces high-quality contract cash flow. The key variable for the next five years is a grander question: will SpaceX become a super-infrastructure company with multiple high-quality cash-flow wheels, or a listed myth defined simultaneously by a great vision and enormous capital consumption.

The conditions that would genuinely make it a better investment target are not complex: the valuation returns to a range that can better accommodate mistakes; Starlink's profit keeps expanding; Starship's commercialization milestones are proven; the AI capex share falls clearly; and debt and refinancing pressure are moderated. Conversely, if over the next two to three quarters the connectivity margin clearly weakens, Starship milestones slip again, and AI losses keep devouring nearly all internal cash flow, then investors can no longer read this as "normal growing pains" and must reexamine whether they have bought a great company's stock far too dear.

For the bull case, I see five core reasons. First, Starlink is already a profit machine in reality, not a concept product. Second, SpaceX's launch and reusability edge has been proven by large-scale data, not a single technical stunt. Third, the government and national-security business gives the company a thicker demand base than pure consumer tech. Fourth, once Starship succeeds, it will further lower costs and open a larger profit pool. Fifth, almost no truly equivalent target exists in the capital markets, and scarcity itself supports a valuation premium.

For the bear case, there are likewise at least five reasons. First, the $1.75 trillion valuation is too front-loaded against the existing financials. Second, AI consolidation has markedly worsened earnings quality and turned the once-clear "rocket + Starlink" story into a complex holding structure. Third, capex and debt are already large enough to alter the equity-return path. Fourth, the governance structure is extremely unfriendly to common shareholders, invisible in good times and quickly priced in bad times. Fifth, Starship still has not crossed the threshold of genuine commercial reusability, and both technical and regulatory risk remain.

In a pre-mortem, the first concrete script for a 50% loss over the next three years would look like this: from the second half of 2026 into 2027, Starship's key tests keep advancing but never stably achieve high-frequency recovery and on-orbit refueling, and the Space segment's loss narrows less than expected; meanwhile AI capex stays high, leaving the 2027 capex/revenue ratio still markedly above that of mature tech-infrastructure companies; the market stops accepting the "burn cash first, then reign" narrative, and the valuation compresses from $1.75 trillion to $900 billion–$1.1 trillion, with the share price most likely halving. The real basis for this script is that the company already showed $10.1 billion of capex and a $1.94 billion operating loss in one quarter, while Starship is still in its test-flight phase.

The second script is more insidious: it is not that Starship fails, but that the quality of Starlink's growth deteriorates. Suppose subscribers keep growing from 2027 but ARPU declines faster than high-value scenarios scale up, and the connectivity margin falls clearly from its current high-30s%; meanwhile Viasat, AST, Kuiper, and traditional telecom partnership plans divert some scenarios, and the market starts to treat Starlink as "a high-capex global operator" rather than "a super-platform cash machine." Then even with revenue still growing, the valuation multiple would converge toward mature communications companies. In that case, profit might not collapse, but the narrative would.

Finally, a restrained but clear research conclusion.

The above rating is not investment advice but a research judgment based on public information and acknowledging that important gaps in the data remain.

Research uncertainties are mainly fourfold. First, the S-1 is the first systematic disclosure, and later accounting may keep adjusting. Second, historical comparability dropped clearly after xAI consolidation. Third, the final offering range, share count, and float are not yet fully clear, leaving the per-share valuation framework unstable. Fourth, key businesses like Starship and direct-to-cell still carry strong "engineering-milestone risk" at this stage.

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