SpaceX IPO: Full Institutional Underwriting Framework

1. Executive Investment Summary

Space Exploration Technologies Corp. presents one of the most strategically significant, technically differentiated, and structurally complex IPO candidates in the global public equity market. The company is not a conventional launch-services company, satellite operator, broadband provider, defense contractor, AI infrastructure platform, or software company. It is an integrated space, connectivity, AI infrastructure, and founder-controlled strategic platform whose valuation must be underwritten by segment, capital intensity, governance, and post-IPO supply rather than by a single revenue multiple.

Updated as of June 3, 2026 for the Form S-1/A filed June 3, 2026. The June 3 amendment, SEC accession 0001628280-26-040364, supersedes the June 1 amendment for IPO economics and converts the report from a pending-price framework into a concrete IPO-terms framework. SpaceX is offering 555,555,555 primary Class A shares at an expected $135.00 per share, implying approximately $75.0 billion of gross base proceeds, $74.4 billion of estimated net proceeds, an 83,333,333-share underwriter option, and $85.7 billion of estimated net proceeds if the option is exercised in full. At the expected price and the March 31 pro forma share count, the base offering implies roughly $1.765 trillion of basic post-offering equity value and approximately 94.5x 2025 consolidated revenue. The new filing also discloses $7.85 of pro forma as-adjusted net tangible book value per share and $127.15 of immediate dilution to new investors, meaning the IPO buyer is paying a very large premium to the historical capital base. The investment view should stay segmented: Connectivity remains the current profit anchor; Space and Starship are strategic but capital-intensive; AI is potentially valuable but still loss-making, power-constrained, legally exposed, and dependent on contract durability; and founder control, lock-up releases, retail/DSP distribution, and post-IPO resale registration now need to be treated as core underwriting variables rather than footnotes.

The June 3 S-1/A materially changes the current report because expected pricing and issuance are now disclosed. The filing states SpaceX is offering 555,555,555 Class A shares at an expected $135.00 per share, with an underwriter option for 83,333,333 additional Class A shares. Estimated net proceeds are $74.4 billion in the base case and $85.7 billion if the option is exercised in full. The company intends to use proceeds for growth strategy, including AI compute infrastructure, launch infrastructure and vehicles, satellite constellation capacity, and general corporate purposes, while management retains broad discretion over allocation.

The report should explicitly describe the $135.00 price as expected or preliminary, not final effective pricing. The preliminary prospectus remains subject to completion, the registration statement has not yet become effective, and the formal underwriting table still leaves the per-share public offering price, underwriting discounts and commissions, and proceeds rows blank. This matters because Reuters correctly pointed to the $135.00 figure, but the filing itself still has not completed every final pricing table.

The new valuation bridge is demanding. At 13.0759 billion basic post-offering shares and $135.00 per share, the implied basic post-offering equity value is approximately $1.765 trillion, or about 94.5x 2025 consolidated revenue of $18.674 billion. If the 83.3 million-share option is exercised, the share count rises to roughly 13.1593 billion and the implied basic equity value rises to approximately $1.776 trillion. This valuation requires investors to pay today for Starlink growth, launch durability, Starship upside, direct-to-device optionality, AI compute utilization, and long-dated non-Earth optionality.

The immediate dilution disclosure should become a headline item. SpaceX estimates pro forma as-adjusted net tangible book value of $7.85 per share after the offering and immediate dilution of $127.15 per share to new investors. New public investors contribute 48.0% of total consideration while owning 4.2% of shares outstanding after the offering; existing investors paid an average $6.48 per share versus the $135.00 IPO price. This is the cleanest S-1-based expression of the public-investor premium.

Post-offering control remains highly concentrated. The June 3 filing shows 7.380 billion Class A shares and 5.696 billion Class B shares outstanding after the base offering, with Class A representing 11.5% of voting power and Class B representing 88.5%. Elon Musk is expected to hold 82.4% of combined voting power after the offering, or 82.3% if the underwriter option is exercised. The beneficial ownership table can show 84.4% because it includes 350.0 million Class B options exercisable within 60 days. The report should reconcile those denominators rather than presenting one control figure as definitive.

The scale of operations is exceptional. Filing graphics disclose approximately 7,400 metric tons delivered to orbit and approximately 10.3 million Starlink subscribers as of March 31, 2026. Starlink subscribers increased from 2.3 million in 2023 to 4.4 million in 2024, 8.9 million in 2025, and 10.3 million in Q1 2026. Falcon and Starship launch counts increased from 98 in 2023 to 138 in 2024 and 170 in 2025. These metrics show real operational scale, not merely an early-stage concept. The company has credible evidence of technical differentiation in launch cadence, orbital payload delivery, constellation deployment, and satellite broadband adoption.

Additional filing detail reinforces the scale: as of March 31, 2026, SpaceX disclosed approximately 9,600 Starlink broadband and mobile satellites in low-Earth orbit, connectivity to approximately 10.3 million Starlink subscribers across 164 countries, territories, and other markets, and a dedicated satellite-to-mobile constellation of approximately 650 V1 Mobile satellites.

The financial profile is far more complicated than the operating metrics imply. Public reports citing the S-1 state that consolidated revenue was $10.387 billion in 2023, $14.015 billion in 2024, and $18.674 billion in 2025. The company reportedly generated a $4.628 billion net loss in 2023, $791 million of net income in 2024, and a $4.937 billion net loss in 2025. Reported revenue growth was approximately 34.8% in 2024 and approximately 33.2% in 2025. However, 2025 net loss represented approximately -26.5% of revenue, and 2025 capital expenditures were $20.737 billion, exceeding annual revenue. This creates a critical distinction between operational scale and equity value creation. Revenue growth is strong, but cash conversion, capital intensity, dilution, and segment profitability remain the primary underwriting issues.

Starlink appears to be the highest-quality economic asset within the company. Reuters reported that Starlink generated approximately $11.387 billion of 2025 revenue and $4.423 billion of operating profit, implying an operating margin of approximately 38.8% before any adjustments for segment allocation, satellite depreciation, launch cost allocation, terminal subsidies, or replacement capex. This is a powerful profit engine and provides the strongest fundamental anchor for valuation. However, Starlink monthly ARPU declined from $99 in 2023 to $91 in 2024, $81 in 2025, and $66 in Q1 2026. That decline is material. It may reflect international expansion, lower-priced geographies, consumer mix, promotional pricing, direct-to-device mix, or broader price optimization. It may also indicate incremental customer monetization pressure. Without cohort-level retention, gross margin, churn, geography, product mix, and terminal subsidy disclosure, Starlink subscriber growth cannot be evaluated as a pure unit-economic positive.

The AI/xAI/X component materially changes the investment case. Reuters reported that the AI division generated approximately $3.201 billion of 2025 revenue, incurred a $6.355 billion operating loss, and accounted for 61% of consolidated 2025 capex. The AI segment therefore appears to be the primary driver of consolidated losses and capital intensity. The company’s AI capacity is large, with filing graphics disclosing nameplate compute draw increasing from 0.3 GW in 2024 to 0.8 GW in 2025 and 1.0 GW in Q1 2026. However, nameplate compute capacity is not equivalent to contracted utilization, durable revenue, gross margin, free cash flow, or return on invested capital. Axios reported that Anthropic agreed to pay SpaceX $1.25 billion per month, or $15 billion annually, through May 2029 for compute resources, but that the arrangement is cancellable by either party with 90 days’ notice. If accurate, that feature materially weakens the backlog quality and should prevent valuation of the revenue stream as long-duration take-or-pay infrastructure revenue.

The company’s balance sheet is large but must be evaluated in the context of aggressive capital deployment. Reuters reported year-end 2025 cash and cash equivalents of $24.747 billion, total assets of $92.079 billion, and total liabilities of $50.754 billion. The company also entered into a $20.0 billion bridge loan in March 2026, with proceeds used to refinance existing subsidiary indebtedness. The credit agreement requires 100% of net cash proceeds from a Qualified IPO to prepay loans within 6 months. This creates a meaningful use-of-proceeds issue. IPO proceeds may be used partly for balance-sheet repair, debt refinancing, and repayment of merger-related or subsidiary indebtedness rather than purely for high-return growth investment.

Supply is more analyzable after the June 3 filing. Founder shares are locked for 366 days with no early release, and founder plus extended-lock-up shares represent approximately 7.8 billion shares, or roughly 60% of the post-offering common share count. Other holders face a staged 180-day release calendar tied partly to the first earnings release and trading-performance thresholds. The directed share program can reserve up to 5% of offered shares, or about 27.8 million base shares, and DSP shares are not subject to lock-up.

The revised investment conclusion is not that the IPO is uninvestable. The assets are too rare for a simple avoidance call. The conclusion is that the current report must stop framing key IPO terms as pending and must underwrite the transaction at the now-disclosed $135.00 expected price, $74.4 billion base net proceeds, $1.765 trillion basic equity value, $127.15 immediate dilution, 88.5% Class B voting power, and staged supply calendar.

2. IPO Thesis And Key Questions

The bull case rests on the possibility that SpaceX is building the dominant global infrastructure layer for orbital transportation, satellite broadband, mobile connectivity, government space services, and AI compute. The company has evidence of real-world scale that most private-market technology companies lack at IPO: approximately 10.3 million Starlink subscribers, approximately 9,600 Starlink broadband and mobile satellites in LEO, high launch cadence, and a Connectivity segment that generated $4.423 billion of 2025 operating profit.

The bear case is now more concrete. The June 3 S-1/A implies approximately $1.765 trillion of basic post-offering equity value at the expected $135.00 IPO price, or about 94.5x 2025 consolidated revenue. New investors would pay $135.00 per share for a company with $7.85 of pro forma as-adjusted net tangible book value per share after the offering. That valuation cannot be justified by conventional aerospace, satellite communications, telecom infrastructure, defense, cloud infrastructure, or software metrics alone.

The central debate is whether SpaceX should be valued as a rare, vertically integrated global infrastructure compounder or as a founder-controlled, capital-intensive option vehicle whose most valuable future businesses remain unproven at scale. The June 3 filing makes the debate sharper because public investors now have a real expected price, share issuance, proceeds estimate, voting split, dilution figure, and lock-up calendar to underwrite.

The 10 questions most likely to determine whether the IPO works as a public equity are as follows:

• 1. What are the pro forma revenue, gross margin, operating margin, depreciation, capex, operating cash flow, and free cash flow by segment for Space, Connectivity, AI, direct-to-device, and government services?
• 2. What explains the decline in Starlink monthly ARPU from $99 in 2023 to $66 in Q1 2026, and how much of the decline is international mix, lower-priced plans, promotions, or structural pricing pressure?
• 3. What portion of AI Solutions & Infrastructure revenue is contracted, external, recurring, non-cancellable, customer-concentrated, related-party-linked, or dependent on a small number of AI customers?
• 4. How much of the $74.4 billion base net proceeds is effectively earmarked for bridge-loan repayment, AI compute, launch infrastructure, satellite constellation growth, spectrum/direct-to-device, working capital, or discretionary M&A?
• 5. What is the normalized free-cash-flow profile of Connectivity after satellite replacement capex, terminal subsidies, support cost, internal launch cost, and spectrum costs?
• 6. What are the economics of Starship after internal launch allocation, refurbishment, regulatory cadence, launch-site capex, and reliability reserves?
• 7. What is the fully diluted share count after preferred conversion, the xAI/X consolidation, RSUs, options, SAUs, warrants, founder restricted shares, ESPP reserves, the 83.3 million-share option, and Form S-8 registration?
• 8. How should investors reconcile the 82.4% post-offering Musk voting-power figure with the 84.4% beneficial ownership table figure that includes 350.0 million Class B options exercisable within 60 days?
• 9. How much supply can become available through the DSP, online brokerage allocation, staged lock-up releases, Form S-8 registration, registration rights, and Rule 144 beginning 90 days after effectiveness?
• 10. What valuation framework justifies paying roughly 94.5x 2025 revenue and $127.15 of immediate dilution per share without clearer segment-level free-cash-flow disclosure?

The report should keep the investment posture as valuation-sensitive engagement. SpaceX is not a generic speculative IPO, but at the expected $135.00 price the investor is underwriting a multi-segment infrastructure platform with founder control, intense capital needs, AI contract uncertainty, legal/regulatory complexity, and a very large public-investor premium.

3. Company Overview

Space Exploration Technologies Corp. was founded in 2002 and has developed from a private launch company into a vertically integrated space, communications, and AI infrastructure platform. The company’s historical milestones include the first privately developed liquid-fuel rocket to reach orbit in 2008, the first private spacecraft to dock with the International Space Station in 2012, reusable orbital-class booster achievements in 2015 and 2017, and private astronaut transportation to orbit and the ISS beginning in 2020. These milestones matter because they establish that the company’s core launch capability is not theoretical. The enterprise has achieved repeated operational milestones in technically difficult markets where barriers to entry are substantial.

The company’s business can be separated into at least 4 economic layers. The 1st is launch and space transportation, including Falcon, Dragon, Starship, launch services, cargo, crew transport, and related mission services. The second is Starlink, a low-Earth-orbit broadband network serving residential, enterprise, mobility, maritime, aviation, government, and direct-to-device use cases. The third is government and national-security services, including secure satellite services and potential Starshield-related capabilities. The fourth is AI/xAI/X, which includes AI compute infrastructure, AI software and model activity, and X-related platform assets following the February 2026 xAI/X reorganization.

Starlink is the current economic center of gravity. Public reports citing the filing state that 2025 revenue included approximately $11.387 billion from Starlink, approximately $4.1 billion from space operations, and approximately $3.201 billion from AI. On that basis, Starlink represented approximately 61% of consolidated 2025 revenue, space operations represented approximately 22%, and AI represented approximately 17%. The relative profitability of those segments is even more important than the revenue split. Starlink reportedly generated $4.423 billion of operating profit, while AI reportedly generated a $6.355 billion operating loss. This suggests that Starlink is the primary cash-generating asset, while AI is the primary investment and loss center.

Starlink’s revenue model is primarily recurring subscription revenue, supplemented by hardware, service-plan, mobility, enterprise, government, maritime, aviation, and direct-to-device revenue streams. The model is not a pure SaaS model because it requires satellites, launches, terminals, ground stations, spectrum coordination, regulatory approvals, customer support, and replacement capex. It has recurring-revenue characteristics, but it is fundamentally a communications infrastructure business with material capital requirements.

Launch and space operations are mission-based and infrastructure-intensive. Revenue is likely driven by commercial customers, government missions, internal Starlink deployments, cargo and crew programs, and other space-transportation services. Internal demand from Starlink is strategically valuable because it creates launch utilization and accelerates cost-curve learning. However, internal launches are not equivalent to external revenue unless transfer pricing and segment economics are disclosed. The filing does not provide enough detail to determine how much launch activity is externally monetized versus internally consumed.

The AI business is the least transparent and most consequential addition. The significant subsidiaries list includes X Corp. and X.AI LLC. The filing graphics describe AI activation, acquisition of X, acquisition by SpaceX, construction of a gigawatt-scale AI training cluster, a gigawatt-scale Megapack battery installation, and nameplate compute draw of 1.0 GW in Q1 2026. These disclosures indicate large-scale infrastructure deployment, but they do not provide sufficient detail on customer contracts, utilization, margins, depreciation, useful lives, power costs, GPU procurement, or long-term return on invested capital.

Geographic exposure is not disclosed with sufficient granularity. This is a meaningful gap because Starlink economics differ materially by region. Developed-market residential broadband, emerging-market broadband, enterprise mobility, maritime, aviation, government, and direct-to-device wholesale connectivity have different pricing, churn, regulation, terminal subsidy, channel, and margin characteristics. A consolidated subscriber count does not provide enough information to evaluate the durability of subscriber economics.

The S-1 revenue disaggregation prevents over-generalization of the business mix. Space revenue contains both Launch Services and Launch & Development. Connectivity contains both Consumer and Enterprise & Government revenue, with Enterprise & Government including Starlink Mobile service offerings. AI contains both Advertising and AI Solutions & Infrastructure. The last distinction matters most: the AI segment should not be valued as if all revenue were external AI compute revenue.

S-1/A comparability caveat: the annual revenue table should be read with the common-control reorganization and segment-realignment caveat. The financial statements present X, xAI, and SpaceX common stock and operating history in a combined structure after the xAI/X transactions, and prior periods conform to the current Space, Connectivity, and AI reporting structure. Historical AI and platform revenue therefore should not be read as clean standalone organic SpaceX history.

4. Business Model Quality

The consolidated business model is not easily categorized. It resembles a hybrid of high-quality recurring connectivity revenue, capital-intensive telecom infrastructure, mission-based aerospace services, defense-adjacent government contracting, AI compute infrastructure, AI software, and platform monetization. This hybrid structure creates substantial strategic optionality but complicates valuation. The company should not be valued as a simple SaaS company, a launch contractor, a telecom carrier, or an AI cloud provider. Each segment has different economics, risks, growth rates, margins, and capital intensity.

Starlink appears to be the highest-quality portion of the business model. It has recurring revenue, large subscriber growth, global reach, potential operating leverage, and a strong strategic moat derived from constellation scale and vertically integrated launch capacity. The reported 2025 operating profit suggests that Starlink can be highly profitable at scale. However, Starlink’s quality depends heavily on metrics that are not fully disclosed. Gross retention, net retention, churn, terminal subsidy, customer acquisition cost, satellite replacement cost, ground-network cost, and geographic mix are all necessary to assess whether the business compounds cash flow or merely scales revenue through heavy capital investment.

The decline in Starlink ARPU is a central quality-of-revenue concern. Subscriber growth from 2.3 million in 2023 to 10.3 million in Q1 2026 is impressive, but ARPU decline from $99 to $66 over the same period indicates that incremental subscribers are monetizing at lower average monthly levels. This may be acceptable if incremental cost to serve declines faster than ARPU, if utilization improves network economics, or if international expansion opens durable large markets. It is problematic if ARPU decline reflects price pressure, competition, lower-income geographies, weaker customer quality, or increasing dependence on lower-margin wholesale arrangements.

Launch operations are strategically advantaged but not necessarily high-visibility recurring revenue. The business has high technical barriers, high fixed costs, regulatory dependency, mission risk, and potential customer concentration. The strategic value lies in vertical integration. SpaceX can lower its own Starlink deployment cost, control launch cadence, and capture external launch demand. That combination creates a flywheel. However, public investors need external launch revenue, internal transfer pricing, launch gross margins, refurbishment costs, failure reserves, and insurance expense to determine whether the launch business is independently attractive or primarily an enabler of Starlink and Starship.

The AI business currently weakens consolidated business-model quality because it appears to be highly capital-intensive and deeply loss-making. AI compute infrastructure can produce valuable revenue when capacity is scarce, customers are committed, and utilization is high. It can also destroy capital if GPU depreciation, power costs, debt financing, customer concentration, and technology replacement cycles outpace contracted revenue. The reported $6.355 billion AI operating loss and 61% share of 2025 consolidated capex indicate that AI has not yet demonstrated high-quality earnings power.

The company therefore has two conflicting business-model attributes. The core launch and Starlink assets appear rare and differentiated. The consolidated enterprise, however, is becoming more capital-intensive, more complex, and less transparent due to AI expansion and the xAI/X integration. This tension will likely define public-market reception. Investors seeking a clean Starlink/launch compounder may object to underwriting large AI and non-Earth compute projects under a founder-controlled governance structure.

The incremental business-model point is that AI revenue and AI capital intensity are mismatched in maturity. Advertising remains the larger AI revenue line in 2025, while AI Solutions & Infrastructure is the line most relevant to compute-infrastructure valuation. At the same time, AI represented the majority of 2025 and Q1 2026 capex. This combination makes AI a high-variance option on utilization, customer durability, model adoption, data-center cost, power access, and GPU depreciation rather than a proven high-quality recurring revenue business.

5. Product And Technology Analysis

The strongest evidence of technical differentiation is in launch, payload delivery, and Starlink constellation deployment. The filing discloses 96 Falcon launches and 2 Starship launches in 2023, 134 Falcon launches and 4 Starship launches in 2024, 165 Falcon launches and 5 Starship launches in 2025, and 40 Falcon launches in Q1 2026. It also discloses 1,210 metric tons delivered to orbit in 2023, 1,699 metric tons in 2024, 2,213 metric tons in 2025, and 556 metric tons in Q1 2026. These are hard operational metrics. They support a conclusion that SpaceX has achieved launch cadence and payload throughput that are highly difficult for competitors to replicate.

Launch productivity appears to be improving modestly. Delivered mass per launch was approximately 12.3 metric tons in 2023, 12.3 metric tons in 2024, 13.0 metric tons in 2025, and 13.9 metric tons in Q1 2026. This metric is influenced by payload mix, mission type, internal Starlink deployment, and Starship testing, but it indicates that the company is increasing aggregate orbital throughput. This matters because orbital mass delivery is a core input into Starlink capacity, Starship economics, and potential non-Earth infrastructure ambitions.

Starlink’s technical differentiation is supported by operating scale. The company disclosed approximately 10.3 million Starlink subscribers as of March 31, 2026. The network’s scale is enabled by vertical integration in launch, satellite manufacturing, ground infrastructure, user terminals, and network operations. This combination is materially different from a satellite operator that must purchase launch capacity from third parties. SpaceX can coordinate launch cadence with constellation capacity needs, optimize satellite design for its own launch economics, and use internal demand to accelerate cost learning.

The Starlink moat is not based on one isolated technology. It is a composite of launch cost advantage, constellation size, spectrum position, ground network, terminal ecosystem, brand, service availability, regulatory footprint, and engineering execution. This composite moat is meaningful. However, it is not absolute. Competitors and substitutes include terrestrial fiber, cable broadband, fixed wireless, mobile networks, Viasat, Hughes/EchoStar, Eutelsat/OneWeb, Amazon Kuiper, AST SpaceMobile, Globalstar, Iridium, government networks, and sovereign satellite systems. In dense urban markets, terrestrial broadband is structurally advantaged on cost and latency. In remote, mobile, maritime, aviation, and underserved markets, Starlink’s relative advantage is stronger.

The AI technology disclosure is less decision-useful than the space disclosure. Nameplate compute draw of 1.0 GW in Q1 2026 demonstrates infrastructure scale but does not prove model differentiation, durable customer demand, training efficiency, inference economics, customer retention, or cash margin. Large AI compute deployments can be valuable if they secure long-term customers and cost-advantaged power. They can be dangerous if they require continuous GPU refresh cycles, debt financing, low-margin contracts, or exposure to a few customers with termination flexibility.

The founder compensation milestones reveal the strategic ambition but also the speculative nature of certain long-term initiatives. The March 2026 restricted-stock award includes a compute milestone requiring completion of non-Earth-based data centers capable of delivering 100 terawatts of compute per year. The January 2026 award includes a Mars Colony Milestone involving a permanent human colony on Mars with at least 1 million inhabitants. These milestones indicate the company’s long-term strategic direction, but they should not be capitalized as near-term economic value without proof of technical feasibility, customer demand, regulatory approval, launch cost, power economics, cooling, repairability, orbital servicing, and insurability.

Key product and technology disclosures remain incomplete. The filing does not provide sufficient detail on patents, cybersecurity certifications, model-performance benchmarks, AI training efficiency, Starlink uptime, latency by geography, terminal failure rates, satellite useful lives, launch failure reserves, refurbishment economics, network congestion, capacity utilization, or direct-to-device technical performance. The public-market underwriting therefore must rely on hard operating outputs where available and apply skepticism to long-dated technology claims that lack economic substantiation.

The S-1 provides specific evidence for Starlink Mobile and AI infrastructure. Starlink Mobile had approximately 650 V1 Mobile satellites and approximately 7.4 million monthly unique devices across approximately 30 countries as of March 31, 2026, with approximately 30 mobile network operator partners across six continents. SpaceX also states that it expects Starship to deploy approximately 50 mobile satellites per launch and that next-generation V2 Mobile satellites are planned for 2027, supported by EchoStar spectrum assets and optimized 5G protocols.

On AI infrastructure, the filing identifies COLOSSUS and COLOSSUS II as the terrestrial compute facilities supporting approximately 1.0 GW of nameplate compute draw as of Q1 2026, but explicitly defines nameplate compute draw as installed GPU capacity multiplied by all-in power draw rather than actual power consumption or utilization. The filing also describes Terafab as a chip-manufacturing initiative with Tesla and Intel targeting 1 terawatt per year of compute hardware over the long term, and says SpaceX expects to begin deploying orbital AI compute satellites as early as 2028. These disclosures should be treated as strategic and technical ambition until contract utilization, cost per token, depreciation, useful lives, power economics, and customer renewal behavior are disclosed.

6. Customers, Demand, And Go-To-Market

Customer demand appears broad in Starlink and more opaque in launch, government services, and AI. Starlink subscriber growth is the clearest evidence of demand. Subscribers increased from 2.3 million in 2023 to 4.4 million in 2024, 8.9 million in 2025, and 10.3 million in Q1 2026. This is rapid scaling for a hardware-enabled global communications service. It suggests strong product-market fit in markets where terrestrial broadband is unavailable, unreliable, expensive, or insufficient for mobile and remote use cases.

The go-to-market model likely differs substantially across Starlink segments. Residential broadband is likely sold directly to consumers through online channels and hardware distribution. Enterprise, maritime, aviation, mobility, and government offerings likely involve longer sales cycles, higher service-level requirements, integration work, and account-based selling. Direct-to-device connectivity likely involves carrier partnerships, wholesale economics, spectrum arrangements, and regulatory approvals. These channels have different margins and retention profiles. The filing does not disclose enough detail to assess segment-level sales productivity.

The Starlink demand signal must be analyzed alongside ARPU compression. A company can add subscribers quickly by entering lower-priced geographies, lowering prices, subsidizing hardware, bundling services, or shifting into wholesale channels. None of those is necessarily negative if incremental contribution margin is attractive. However, the absence of churn, cohort ARPU, terminal subsidy, and contribution margin disclosures prevents a clean evaluation of whether demand is high-quality or increasingly price-sensitive.

Launch demand is evidenced by launch count and orbital mass delivered, but customer-level demand is not sufficiently disclosed. The filing graphics provide total launches and mass to orbit, but not external versus internal launch mix, customer concentration, backlog, average mission revenue, launch margin, government versus commercial mix, or contract duration. Internal demand from Starlink is strategically valuable because it supports utilization and cost learning. However, for valuation purposes, external customer demand and mission profitability are needed to underwrite launch as a standalone business.

Government and national-security demand may be strategically significant, but it is not disclosed with enough granularity to quantify. SpaceX’s launch and secure satellite services likely have important government relevance. Starshield-related services may increase national-security exposure. This can create durable demand, high barriers to entry, and strategic importance. It can also create classified disclosure limitations, procurement risk, political risk, export-control exposure, and customer concentration.

AI demand is potentially large but high-risk. Axios reported an Anthropic compute arrangement worth $1.25 billion per month through May 2029, but cancellable by either party with 90 days’ notice. That is an important distinction. A contract of that size would dramatically improve near-term revenue visibility if it is utilized and profitable. The termination flexibility, however, makes it materially less valuable than non-cancellable backlog. AI demand must be evaluated based on committed duration, termination rights, utilization, customer concentration, price escalators, power pass-throughs, depreciation, and renewal economics.

Customer concentration is not disclosed with sufficient granularity. This is a major gap. The most important missing disclosures include top 10 customers by segment, government revenue share, AI customer concentration, Starlink enterprise concentration, direct-to-device partner exposure, launch backlog concentration, and related-party revenue. For a company with a large AI compute business and major government exposure, customer concentration could materially affect valuation.

Customer concentration is material. The S-1 reports that Customer A represented 20.9% of consolidated revenue in 2025, 24.2% in 2024, and 25.2% in 2023, with revenue from that customer relating to all three segments. That is still meaningful concentration for an issuer seeking a very large public-market valuation. The identity, contract duration, renewal terms, segment exposure, and related-party status of this customer are important diligence items.

Starlink Mobile also changes the go-to-market discussion. The S-1 says SpaceX charges mobile network operators either fixed fees or earns revenue through revenue-sharing arrangements based on connectivity services included in partner plans. That structure can scale quickly through carrier distribution, but it may carry lower direct customer ownership, more partner bargaining power, and different ARPU/margin characteristics than consumer broadband.

7. Market Opportunity And TAM Credibility

The market opportunity remains exceptionally large in narrative terms, but the June 3 S-1/A requires stricter TAM language. The filing relies on third-party industry reports that SpaceX did not commission and on management estimates and calculations. It cautions investors not to give undue weight to market-size estimates. The report should therefore stop using very large TAM references as valuation support unless they are tied to monetizable revenue, contract duration, pricing, margins, capex, and regulatory feasibility.

Starlink’s relevant addressable market is not simply global internet spending. The actual monetizable opportunity is constrained by affordability, capacity density, spectrum rights, service quality, regulatory access, terminal cost, and terrestrial substitutes. In remote and underserved markets, Starlink may have a very large advantage. In dense urban markets, fiber, cable, and fixed wireless may be lower-cost and better suited to high-density demand. In emerging markets, demand may be large but ARPU may be structurally lower. The ARPU decline already indicates that incremental addressable demand is not uniformly priced at early-adopter levels.

Direct-to-device connectivity expands the addressable market beyond fixed broadband, but the economics may be meaningfully different. Mobile dead-zone coverage, emergency connectivity, rural mobile service, and carrier wholesale arrangements can create large strategic value. However, wholesale telecom economics may involve lower ARPU, carrier bargaining power, spectrum constraints, regulatory conditions, and lower direct customer ownership. Direct-to-device should therefore be valued separately from residential Starlink broadband.

Launch TAM is also more constrained than popular space-economy narratives imply. Lower cost to orbit can expand demand by enabling new satellite, manufacturing, defense, science, and orbital infrastructure applications. However, payload demand, regulatory cadence, insurance, customer budgets, mission reliability, and government procurement cycles still matter. SpaceX can expand the market, but the market will not automatically absorb unlimited launch capacity at attractive margins.

The AI TAM is the largest and least disciplined part of the market story. AI compute demand is real and substantial, but it is also competitive, power-constrained, capital-intensive, and customer-credit-sensitive. The June 3 filing supports a large AI infrastructure ambition, including terrestrial compute buildout and long-dated non-Earth compute concepts, but investors should value it as option value until contract quality, utilization, gross margin, power cost, data-center permitting, and customer concentration are disclosed.

The TAM for non-Earth compute and Mars-related infrastructure should remain speculative option value. The S-1/A links founder incentives to permanent Mars colony milestones and non-Earth data centers capable of 100 terawatts of compute per year, but current disclosure does not show commercial revenue, regulatory structure, maintenance feasibility, customer contracts, or unit economics for those markets. The report should frame these as long-dated strategic options, not near-term underwriting anchors.

8. Industry Structure And Competitive Position

SpaceX appears structurally advantaged in launch. The combination of reusable launch vehicles, high launch cadence, vertical manufacturing, internal Starlink demand, and demonstrated payload throughput creates a formidable barrier to entry. Competitors can challenge SpaceX in specific niches, sovereign programs, responsive launch, small launch, government procurement categories, and specialized missions, but matching SpaceX’s integrated scale appears difficult. Relevant launch and space-infrastructure competitors include Blue Origin, United Launch Alliance, Arianespace, Rocket Lab, national space agencies, sovereign launch providers, and emerging launch companies.

Rocket Lab is the most relevant public space-infrastructure and launch comparable, but it is far smaller. Rocket Lab reported Q1 2026 revenue of $200.3 million, GAAP gross margin of 38.2%, and backlog of $2.2 billion. It is useful for understanding public-market appetite for space infrastructure, but it is not a direct valuation comp. SpaceX has far greater launch scale, Starlink revenue, internal payload demand, and AI exposure. Rocket Lab’s financials also provide a reminder that space companies can grow rapidly while remaining much smaller and less diversified than SpaceX.

Starlink’s competitive position is strong but not unconstrained. Its advantages include constellation scale, low-Earth-orbit architecture, launch integration, brand, service availability, and increasing ecosystem reach. Competitors include Viasat, Hughes/EchoStar, Eutelsat/OneWeb, Amazon Kuiper, Iridium, Globalstar, AST SpaceMobile, terrestrial broadband, fixed wireless, mobile networks, and government-backed systems. Amazon Kuiper is a particularly important potential competitor because Amazon has capital, cloud infrastructure, device distribution, and customer relationships. Starlink’s head start is meaningful, but competition can pressure pricing, spectrum access, hardware economics, and regulatory approvals.

Satellite communications market structure is shaped heavily by regulation. The FCC reportedly approved 7,500 additional Gen2 Starlink satellites while permitting SpaceX to operate up to 15,000 satellites under the relevant framework, with deployment milestones extending into 2028 and 2031. Regulatory approvals can serve as both a moat and a bottleneck. They limit fast replication by competitors, but they also expose SpaceX to spectrum disputes, orbital-debris concerns, international licensing, national-security review, sanctions, launch approvals, and environmental review.

The direct-to-device market is likely to become more competitive and more partner-dependent. Reuters reported that Verizon, AT&T, and T-Mobile formed a joint venture to address mobile dead zones, while the FCC approved spectrum-related transactions involving SpaceX, EchoStar, and AT&T. SpaceX is reportedly investing approximately $17 billion in Starlink direct-to-device. This market can be large, but it is unlikely to behave like a high-ARPU direct-consumer broadband market. Carrier relationships, spectrum economics, wholesale pricing, and regulatory conditions will be central.

The AI infrastructure market is structurally more competitive than launch and Starlink. SpaceX competes with hyperscalers, specialized AI cloud providers, model companies, GPU owners, data-center operators, and enterprise AI platforms. Relevant competitors and adjacent players include Microsoft, Amazon, Google, Meta, Oracle, CoreWeave, OpenAI infrastructure partners, Anthropic’s own infrastructure arrangements, NVIDIA-linked cloud partners, and other large-scale AI infrastructure providers. AI compute is attractive when capacity is scarce, but competitive advantage can erode quickly if supply expands, utilization falls, or customers internalize capacity.

CoreWeave is a relevant public AI infrastructure comp but not a clean one. CoreWeave has large revenue backlog, significant capex requirements, meaningful debt, and customer concentration concerns. It is useful for analyzing the economics of AI cloud infrastructure, but SpaceX differs because AI is embedded within a much broader space, communications, government, and platform business. The comparison highlights the key risk: AI infrastructure can create very high growth and very high valuation, but also very high debt, capex, and customer-concentration risk.

9. Financial Statement Analysis

Reported consolidated financials show strong revenue growth but volatile profitability and heavy capital intensity. Revenue reportedly increased from $10.387 billion in 2023 to $14.015 billion in 2024 and $18.674 billion in 2025. Net income reportedly moved from a $4.628 billion loss in 2023 to $791 million of income in 2024 and then to a $4.937 billion loss in 2025. The company therefore demonstrated an ability to reach profitability in 2024, but that profitability did not persist after the expansion of AI and related capital-intensive initiatives.

The revenue trajectory appears attractive on a standalone basis. Growth of approximately 34.8% in 2024 and approximately 33.2% in 2025 is strong for a company with this level of revenue scale. However, growth quality is uneven. Starlink growth appears volume-driven and tied to subscriber adoption. Launch growth appears tied to launch cadence and payload delivery, with an unknown mix of internal and external demand. AI growth may be driven by compute infrastructure scale and customer demand, but the reported operating losses and capex intensity indicate that revenue growth has not yet translated into earnings quality.

The income statement is likely distorted by segment mix changes. Starlink reportedly generated $4.423 billion of operating profit in 2025, while AI reportedly generated a $6.355 billion operating loss. If those figures are accurate, consolidated performance is the result of a profitable Starlink business offset by an aggressive AI investment cycle. The 2025 consolidated loss should not be interpreted as evidence that the entire enterprise has weak unit economics. It should be interpreted as evidence that segment-level disclosure is indispensable.

The balance sheet is large but leveraged to continued access to capital. Reported year-end 2025 cash of $24.747 billion appears substantial, but it must be evaluated against $20.737 billion of 2025 capex, the $20.0 billion bridge loan, AI infrastructure needs, Starlink satellite replenishment, Starship development, direct-to-device investment, potential spectrum costs, potential acquisition obligations, and working-capital demands. A large cash balance can be consumed quickly when annual capex exceeds revenue.

The $20.0 billion bridge loan is a key financial statement and IPO-structure issue. The agreement provides for an aggregate $20.0 billion term loan facility, with proceeds used to refinance existing subsidiary indebtedness. The facility includes leading banks as arrangers and requires prepayment with 100% of net cash proceeds from a Qualified IPO within 6 months. This means that IPO proceeds may be substantially directed to debt repayment if the facility remains outstanding. Public investors should not assume that all primary proceeds will be available for incremental growth investment.

The company’s capex profile is the most important financial statement concern. Reported 2025 capex of $20.737 billion exceeded reported 2025 revenue of $18.674 billion. AI reportedly accounted for 61% of consolidated capex. This level of capex can be rational if it creates high-utilization, high-margin, long-duration infrastructure revenue. It is value-destructive if demand is cancellable, margins are low, useful lives are short, or capacity becomes commoditized.

Free cash flow is likely the most important missing metric. The reported net loss and capex intensity suggest that free cash flow was materially negative in 2025 absent extraordinary working-capital inflows or other financing effects. For valuation, free cash flow matters more than adjusted EBITDA or even operating income. Satellite broadband, launch infrastructure, AI compute, and data centers require sustained capital investment. Earnings quality should be evaluated after capex, depreciation assumptions, replacement cycles, and working capital.

The auditor consent references a reorganization under common control, a stock split, and a change in reportable segments shortly before the filing. This is important. Common-control reorganizations can affect historical presentation, segment reporting, related-party disclosures, equity structure, and pro forma comparability. The segment reporting change dated shortly before filing creates a need for reconciled historical segment data. Without that reconciliation, public investors may struggle to compare 2023, 2024, 2025, and 2026 performance.

The June 3 S-1/A adds a concrete pro forma balance-sheet bridge. After the offering, SpaceX estimates $90.305 billion of pro forma as-adjusted cash and cash equivalents, $29.111 billion of total long-term debt, and $2.069 billion of spectrum credit agreement and other long-term commitments. This changes the liquidity discussion: the company would have enormous cash resources after the IPO, but that cash must be evaluated against bridge-loan repayment, AI compute buildout, launch infrastructure, satellite constellation scale, spectrum/direct-to-device commitments, and ongoing legal/regulatory obligations.

Several financial statement disclosures are not sufficiently granular in the reviewed materials. Stock-based compensation expense, depreciation by asset class, satellite useful lives, GPU useful lives, terminal subsidies, capitalized software, warranty reserves, deferred revenue roll-forward, contract assets, customer concentration, non-GAAP reconciliations, and segment-level capex are all required for a complete institutional underwriting. Their absence matters because valuation depends on cash conversion and segment returns, not merely revenue scale.

The S-1 segment table is central because it is the cleanest way to understand the consolidated financial profile. Connectivity is the current profit engine, Space is strategically essential but loss-making on a GAAP operating basis, and AI is the dominant loss and capital-consumption center.

Cash-flow analysis should also be separated from earnings. The company generated $6.785 billion of operating cash flow in 2025, but PPE purchases were $20.737 billion. A simple operating cash flow less PPE purchases proxy was therefore approximately $(13.952) billion in 2025 and approximately $(9.060) billion in Q1 2026. This does not replace a full free-cash-flow calculation, but it frames the capital intensity more clearly than net income or Adjusted EBITDA alone.

The S-1/A makes the financing wall explicit. At March 31, 2026, SpaceX had $29.132 billion of aggregate principal debt, $30.286 billion of debt and finance leases, $1.538 billion current debt and finance leases, and scheduled 2027 debt principal maturities of $21.540 billion. The bridge-loan refinancing created a $1.526 billion Q1 2026 extinguishment loss. This should be visible in the financial tables because it affects how investors interpret IPO proceeds, cash balances, and free-cash-flow capacity.

10. Unit Economics And KPI Quality

The disclosed KPI set is useful but incomplete. The company provides mass delivered to orbit, launch count, Starlink subscribers, Starlink ARPU, and nameplate compute draw. These metrics measure physical scale, network adoption, and infrastructure capacity. They do not measure retention, gross margin, free cash flow, return on capital, customer concentration, utilization, or contract durability.

Launch KPIs show real productivity. Total launches increased from 98 in 2023 to 138 in 2024 and 170 in 2025. Mass delivered to orbit increased from 1,210 metric tons in 2023 to 1,699 metric tons in 2024 and 2,213 metric tons in 2025. This supports the view that SpaceX has a meaningful launch cadence and payload-throughput advantage. However, unit economics require more detail. Cost per launch, cost per kilogram, external revenue per launch, internal transfer price, refurbishment cost, launch-failure reserve, and insurance cost are not disclosed with sufficient granularity.

Starlink subscriber KPIs are impressive but incomplete. Subscriber growth from 2.3 million in 2023 to 10.3 million in Q1 2026 demonstrates adoption. ARPU decline from $99 to $66 over the same period introduces a material analytical caveat. The combination of rapid volume growth and declining ARPU can be attractive if cost to serve falls and contribution margin expands. It can be unattractive if incremental customers require higher subsidies, lower prices, higher support costs, or lower-capacity utilization.

Starlink unit economics should be evaluated at the cohort level. Required metrics include subscriber acquisition cost, terminal hardware cost, terminal subsidy, installation cost, churn, gross retention, net retention, monthly ARPU by cohort, usage by cohort, support cost, satellite capacity cost, ground-network cost, launch cost allocation, and contribution margin. None of these metrics is disclosed in sufficient detail. Without them, subscriber growth cannot be translated into lifetime value or payback.

The AI compute KPI of nameplate compute draw is particularly incomplete. Nameplate compute draw increased from 0.3 GW in 2024 to 0.8 GW in 2025 and 1.0 GW in Q1 2026. This indicates infrastructure scale. It does not indicate customer utilization, revenue per MW, gross margin per MW, power cost, uptime, GPU age, depreciation, customer concentration, contract duration, or renewal probability. AI infrastructure companies can show impressive capacity growth while generating poor returns if utilization or pricing disappoints.

The S-1/A makes total launch count less useful unless split between monetized customer launches and internal launches. Falcon launches include a large and rising internal component, while all Starship launches to date are classified as internal. The KPI table should therefore pair launch cadence with the external/internal split.

ARR, net revenue retention, gross retention, churn, CAC, LTV, payback period, RPO, contracted backlog, pipeline conversion, customer concentration, cohort contribution, and average contract value are not disclosed with sufficient granularity. This is a significant disclosure gap. The company is large enough and complex enough that public investors should expect more decision-useful operating metrics by segment.

S-1 KPI definitions change how subscriber, ARPU, and compute disclosures should be interpreted. In particular, Starlink Subscriber count is not the same as end users or enterprise/government customers, Starlink Mobile monthly unique devices is a separate metric, and nameplate compute draw is capacity rather than utilization.

11. Growth Quality

Growth quality is mixed. Starlink growth appears genuine and customer-driven. Launch growth appears operationally strong. AI growth appears potentially large but capital-intensive and loss-making. Consolidated revenue growth of more than 30% at nearly $19 billion of revenue is attractive, but the composition of that growth matters more than the consolidated rate.

Starlink growth is likely organic and volume-driven, but the ARPU decline complicates interpretation. The company is adding subscribers rapidly, but average monthly revenue per subscriber is falling. If the decline reflects entry into large lower-priced markets with strong contribution margins, growth quality may remain high. If the decline reflects price pressure or customer mix deterioration, growth quality may be lower than subscriber numbers suggest. Disclosure does not yet permit a definitive conclusion.

Launch growth is partly a function of internal ecosystem demand. Starlink deployment creates structural launch demand, which improves cadence, learning, and cost absorption. That is a strategic advantage. However, from a public equity perspective, internal demand can obscure the profitability of launch as a commercial service. External backlog, third-party launch revenue, government revenue, and internal transfer-pricing policies are necessary to evaluate quality.

AI growth is the highest-beta growth source. It can materially increase revenue if large customer contracts are signed and capacity utilization is high. It can also reduce earnings quality if contracts are short-duration, cancellable, customer-concentrated, low-margin, or debt-funded. The reported Anthropic arrangement would be highly material if durable and profitable, but the reported 90-day termination feature reduces visibility.

The growth trajectory is therefore not a simple growth-equity story. It is a capital allocation story. The company has a profitable growth engine in Starlink, a strategic and technically advantaged launch business, and a large AI build-out that may either become a major profit pool or consume disproportionate capital. The quality of growth should be judged by capex-to-revenue, segment free cash flow, ARPU stabilization, contracted AI utilization, and return on invested capital.

Growth quality requires separating Connectivity growth, AI growth, and platform/advertising growth because each deserves a different valuation framework. Consumer Starlink revenue growth is supported by subscriber adoption. Enterprise & Government revenue growth is supported by managed connectivity, Starlink Mobile, and government demand. AI segment revenue growth includes both Advertising and AI Solutions & Infrastructure, while the capex profile is overwhelmingly infrastructure-driven. This mix makes headline AI revenue less useful than contracted compute revenue, utilization, gross margin, depreciation, and customer concentration.

12. Margin Structure And Long-Term Earnings Power

Starlink appears to have the clearest long-term earnings power. Reported 2025 operating profit of $4.423 billion on approximately $11.387 billion of revenue implies an operating margin of approximately 38.8%. That is an attractive margin for a communications infrastructure business if it is calculated after appropriate satellite, launch, terminal, and network cost allocations. The key diligence issue is whether this margin reflects full economic cost, including replacement capex and internal launch economics.

The Starlink margin structure could improve with scale if fixed network costs are absorbed across more subscribers, satellite manufacturing costs fall, launch costs decline, and terminal economics improve. It could deteriorate if ARPU continues to decline, capacity becomes congested, terminal subsidies increase, regulatory costs rise, or direct-to-device wholesale economics dilute margins. The balance between these forces is not disclosed in sufficient detail.

Launch margins are harder to assess. The business likely benefits from reuse, cadence, vertical manufacturing, and internal demand. It also carries high fixed costs, development expense, regulatory risk, launch-site capex, refurbishment costs, failure risk, and mission insurance. Starship could materially improve long-term margins if it achieves reliable reuse and lower cost per kilogram. However, Starship remains a technical, regulatory, and capital investment variable rather than a fully proven margin-expansion driver.

AI margins are currently unfavorable based on reported operating losses. AI compute infrastructure can become profitable if contracts are long-term, utilization is high, power is cost-effective, depreciation is manageable, and customers renew. It can remain structurally challenged if capex requirements, GPU replacement cycles, networking cost, debt cost, and customer bargaining power consume economics. The reported $6.355 billion AI operating loss indicates that AI is not yet a proven margin contributor.

Consolidated long-term earnings power depends on whether Starlink remains profitable enough to fund growth while AI moves toward cash profitability and Starship reduces infrastructure cost. If Starlink sustains a high operating margin and AI losses narrow, consolidated earnings could inflect rapidly. If AI capex and losses persist while Starlink ARPU declines, consolidated margins may remain structurally weak despite revenue growth.

Traditional adjusted EBITDA analysis is inadequate for this issuer. Depreciation and capex are economically central. Satellites, launch infrastructure, AI clusters, GPUs, data centers, ground stations, and terminals have real replacement cycles. Free cash flow and return on invested capital are more important than adjusted EBITDA.

The Segment Adjusted EBITDA figures are directionally useful but insufficient. The S-1 reports 2025 Connectivity Segment Adjusted EBITDA of $7.168 billion and operating income of $4.423 billion, but for Space and AI the difference between adjusted measures and GAAP operating losses reinforces the need to focus on asset replacement cycles, Starship development, GPU depreciation, data-center power, and ongoing capex rather than adjusted profit metrics alone.

13. Cash Burn, Liquidity, And Funding Need

The June 3 S-1/A materially changes the funding analysis because the offering size is now disclosed in expected terms. SpaceX is offering 555,555,555 Class A shares at an expected $135.00 per share, generating estimated net proceeds of $74.4 billion in the base offering and $85.7 billion if the underwriter option is exercised in full. This is one of the largest primary financing events ever proposed, but it should not be treated as excess cash. It is capital for an unusually large, unusually capital-intensive platform.

The use-of-proceeds language should be read as a capital-allocation waterfall rather than a generic growth statement. The filing names expansion of AI compute infrastructure, launch infrastructure and launch vehicles, satellite constellation scale and capacity, and general corporate purposes. It also gives management broad discretion. The report should connect that discretion to the $20.0 billion bridge loan, 2027 debt maturity wall, EchoStar spectrum transaction, AI capex, Starship investment, and post-IPO legal/regulatory costs.

The $20.0 billion bridge facility remains central. The facility matures September 2, 2027, with two optional 3-month extensions, and IPO net proceeds must be used to prepay amounts outstanding within 6 months. Even after a large IPO, investors should separate refinancing capital from growth capital. The right diligence question is not whether SpaceX has enough cash after the IPO; it is how much of that cash can produce attractive incremental returns after required and quasi-required uses.

14. Valuation Framework

The valuation framework must now start with the June 3 S-1/A expected IPO price rather than external valuation reports. SpaceX expects to sell 555,555,555 Class A shares at $135.00 per share. Based on 7.380 billion Class A shares and 5.696 billion Class B shares outstanding after the base offering, basic post-offering shares are approximately 13.0759 billion. The expected price therefore implies approximately $1.765 trillion of basic post-offering equity value, or about 94.5x 2025 consolidated revenue of $18.674 billion.

This valuation cannot be underwritten with a single revenue multiple. Connectivity is the current profit anchor and should be valued around subscriber growth, ARPU, retention, replacement capex, and contribution margin. Space and Starship deserve strategic value but require proof of reliability, cadence, external revenue, and cost reduction. AI is the swing factor but requires contract-quality, utilization, power-cost, and gross-margin evidence. Non-Earth compute, Mars-linked milestones, and Kardashev-scale claims should remain option value.

The base-case recommendation remains valuation-sensitive engagement. Participation at $135.00 requires confidence that Connectivity can compound, Starship can lower launch cost and expand demand, AI infrastructure can earn durable high returns, and management can allocate $74.4 billion of proceeds without value leakage through related-party, speculative, or founder-preference projects.

15. Public Comparables

No public comparable is fully adequate. SpaceX combines launch, satellite broadband, government space services, AI infrastructure, AI software, and platform assets under a controlled-company governance structure. The result is a company that does not fit cleanly into aerospace, telecom, software, cloud infrastructure, or defense-sector valuation frameworks.

Rocket Lab is the closest public launch and space-systems comparable. It is relevant because it provides public-market evidence of investor appetite for space infrastructure and launch growth. However, Rocket Lab is much smaller, with Q1 2026 revenue of $200.3 million and backlog of $2.2 billion. It lacks Starlink’s subscriber base, SpaceX’s launch cadence, Starship optionality, and AI/xAI/X exposure. Rocket Lab is useful directionally, not mechanically.

AST SpaceMobile is relevant for direct-to-device satellite connectivity. It captures investor interest in satellite-to-mobile optionality, but it is earlier-stage and far smaller. Its revenue and profitability profile are not comparable to Starlink. ASTS may be relevant for assessing market appetite for direct-to-device narratives, but not for valuing SpaceX’s current earnings power.

Iridium is relevant for satellite-communications durability and service revenue. It has established satellite service revenue and high EBITDA conversion, but its growth rate, constellation architecture, customer base, scale, and Starlink overlap are limited. Iridium is a useful mature satellite communications reference, not a high-growth Starlink comp.

Viasat and Globalstar are relevant to satellite communications but not directly comparable. Their businesses are more legacy satellite-communications-oriented and lack SpaceX’s vertical launch integration. They are useful for assessing satellite-sector cyclicality, capex, regulatory issues, and spectrum value, but not for directly valuing Starlink’s growth profile.

CoreWeave is the most relevant AI infrastructure comp. It highlights the market’s willingness to value large-scale AI compute businesses, but it also highlights debt, capex, customer concentration, and backlog-quality risk. SpaceX’s AI business could be compared to CoreWeave only at the segment level. Consolidated SpaceX contains launch, Starlink, government services, and AI, making direct comparison inappropriate.

Hyperscalers are relevant for AI capex and infrastructure economics but not as direct comps. Microsoft, Amazon, Google, Meta, Oracle, and other large cloud platforms have diversified operating cash flows, enterprise relationships, and established cloud ecosystems. SpaceX may have unique infrastructure and strategic advantages, but it does not yet disclose AI economics comparable to the hyperscalers’ mature cloud businesses.

Tesla is not a direct business comp but may be relevant to market behavior. Founder premium, retail demand, narrative strength, tolerance for long-dated optionality, and governance structure could affect SpaceX’s trading dynamics. However, Tesla comparison can also be misleading because SpaceX has a different capital structure, business mix, customer base, regulatory profile, and ownership architecture.

16. IPO Structure, Dilution, And Capitalization

The June 3 S-1/A requires a full replacement of the IPO-structure section. Price, primary share count, base net proceeds, option share count, voting split, net tangible book value, and immediate dilution are now disclosed in expected terms. The report should replace the prior pending-price caveat with the more precise statement that expected cover-page economics are disclosed, while the formal underwriting price/discount/proceeds table remains incomplete and the registration statement is still preliminary.

The report should also reconcile denominators. Summary/offering tables use March 31 assumptions and show 7.380 billion Class A shares and 5.696 billion Class B shares after the base offering. Beneficial-ownership tables use a May 1 denominator after Class C reclassification, preferred conversion, stock split, and no option, showing 7.488 billion Class A shares and 5.603 billion Class B shares. Mixing those denominators can create apparent inconsistencies in ownership, voting power, and dilution.

17. Ownership, Governance, And Control

Governance remains a core valuation issue after the June 3 S-1/A. SpaceX will be a controlled company. Class B shares carry 10 votes per share, and after the base offering Class B holders will control 88.5% of voting power. Public Class A investors receive economic exposure but very limited governance leverage.

Elon Musk control should be presented with denominator reconciliation. The cover-page and offering disclosure show Musk with 82.4% combined voting power after the base offering and 82.3% if the underwriter option is exercised in full. The beneficial ownership table can show 84.4% because it includes 350.0 million Class B options exercisable within 60 days. Both figures are sourced; they answer different denominator questions.

Musk beneficial ownership includes 849.494 million Class A shares and 5.569 billion Class B shares in the beneficial ownership table, including 350.0 million Class B options exercisable within 60 days. The cover-page post-offering presentation shows 849.494 million Class A shares and 5.219 billion Class B shares. He also holds 91.6% of Class B in the post-offering ownership table and 1.302 billion restricted Class B shares that may be voted while unvested.

Controlled-company status matters. SpaceX intends to rely on Nasdaq and Nasdaq Texas controlled-company exemptions from requirements for a majority-independent board, independent nominating committee, independent compensation committee, and annual committee evaluations. Audit committee independence still applies. The report should describe these exemptions explicitly because minority public investors will be structurally dependent on board process, disclosure, and reputation rather than voting leverage.

For report language, use the exact formulation that SpaceX will be a controlled company after the IPO and intends to rely on controlled company exemptions, while the audit committee remains subject to independence requirements.

18. Management And Board Quality

The June 3 S-1/A increases the importance of management and board-process diligence because SpaceX intends to rely on controlled-company exemptions. Public investors should underwrite the quality of board oversight, related-party review, compensation approval, and capital-allocation process directly, since Class A holders will have limited voting power after the offering.

Management has demonstrated exceptional operational execution in launch, satellite deployment, and Starlink scaling. The disclosed launch cadence, mass-to-orbit metrics, and subscriber growth support a conclusion that the company has unusual engineering and operational capability. Few companies have demonstrated comparable ability to execute across rockets, spacecraft, satellite manufacturing, broadband services, and large-scale infrastructure.

Elon Musk’s role is both a major asset and a major risk. The company’s strategy, control architecture, compensation structure, and long-term ambitions are closely tied to him. Founder-led execution has historically produced extraordinary operational outcomes. However, the same structure creates key-person risk, governance risk, related-party risk, strategic-priority risk, and potential public-market volatility.

Gwynne Shotwell’s role is important for institutional comfort. Reuters reported that President and COO Gwynne Shotwell received $85.8 million in total compensation last year. CFO Bret Johnsen reportedly received $9.8 million. Shotwell’s long-standing operational leadership provides evidence that SpaceX is not solely dependent on founder-level vision. However, the filing materials reviewed do not provide enough detail on full management bench depth, succession planning, public-company readiness, executive turnover, or board committee independence.

Public-company readiness is a separate issue from operational capability. The company is approaching the public market following a major common-control reorganization, xAI/X integration, reportable segment changes, a large bridge loan, aggressive founder compensation awards, and a multi-class control structure. These factors increase the importance of audit infrastructure, internal controls, segment reporting, related-party review processes, and independent board governance.

Board quality cannot be fully assessed without complete board composition, committee membership, independence analysis, audit committee expertise, related-party transaction committee authority, and compensation committee independence. These are particularly important because the company has founder control, major founder awards, xAI/X integration, potential conflicts involving Musk-affiliated entities, and large capex commitments.

19. Accounting Quality And Disclosure Quality

Accounting quality is difficult to fully assess from the currently available public materials because several essential disclosures are incomplete. The most important issue is comparability. PwC’s consent references a reorganization under common control, a stock split, and a change in reportable segments shortly before the filing. These events can materially affect historical financial presentation and segment comparability.

The xAI merger creates significant accounting complexity. The merger agreement includes exchange ratios, treatment of preferred stock, treatment of options and RSUs, cash elections for certain service providers, and tax-free reorganization intent under Section 368(a). The accounting implications include pro forma consolidation, equity classification, stock-based compensation, fair-value measurement, deferred taxes, and segment reporting. Public investors need a complete bridge from standalone SpaceX to the combined SpaceX/xAI/X entity.

Stock-based compensation requires close scrutiny. The founder awards are outside the 2024 Equity Incentive Plan but have terms substantially similar to the plan and carry voting rights from the grant date. Vested xAI options were subject to cash-out based on the excess of $75.46 per share over the option exercise price, while other equity awards were converted into SpaceX-linked awards. These arrangements can create material compensation expense, dilution, and non-recurring cash effects.

Revenue recognition is a major disclosure gap. The company likely has multiple revenue-recognition models: Starlink subscriptions, hardware sales, enterprise contracts, government contracts, launch services, milestone-based programs, AI compute contracts, platform revenue, and possibly software/API revenue. Each has different recognition timing, deferred revenue, contract asset, cancellation, and margin implications. The reviewed materials do not provide sufficient revenue-recognition detail by segment.

Depreciation and useful-life assumptions are central to earnings quality. Satellites, launch infrastructure, ground stations, terminals, GPUs, data centers, batteries, and software systems all require useful-life estimates. Small changes in depreciation lives can materially affect reported operating profit in a capital-intensive business. AI GPU useful lives are especially important because technological obsolescence can occur rapidly.

Non-GAAP metrics are not available in sufficient detail for evaluation. Any adjusted EBITDA, adjusted operating income, or non-GAAP profitability metric should be treated cautiously unless reconciled clearly to GAAP and accompanied by capex and stock-compensation detail. For this company, stock-based compensation, depreciation, and capex are economically meaningful and should not be dismissed as non-cash or non-operating.

20. Risk Factor Analysis

The June 3 S-1/A moves several risks from generic IPO disclosure to valuation-critical underwriting variables. The key risks are no longer just execution and valuation; they include immediate dilution, founder control, staged supply, AI contract quality, legal/regulatory exposure, cyber/classified-data risk, uninsured launch and satellite losses, bridge-loan repayment, and capital-allocation discretion over a $74.4 billion proceeds pool.

The AI/X risk language should be expanded. The filing describes regulatory, litigation, and platform exposure tied to Grok and X, including nonconsensual explicit imagery, content involving women and children in sexualized contexts, the Irish Data Protection Commission inquiry launched in February 2026 regarding EU personal data including children and Grok within X, FTC chatbot safety inquiries, app-store rules, IP claims, and content moderation issues. These are not ordinary aerospace risks; they are platform, AI safety, privacy, and distribution risks embedded inside the public company.

Cybersecurity risk deserves separate treatment because SpaceX handles government, national-security, and classified or government-controlled information. The filing identifies nation-state, state-sponsored, cybercriminal, insider, and AI-enhanced threats. For a company with launch, satellite, defense, AI, and communications infrastructure, a breach is not merely an IT cost; it can affect government trust, classified contracting, export-control posture, uptime, customer commitments, and national-security positioning.

Asset-loss risk should be made more explicit. SpaceX does not typically obtain insurance for satellites, payloads, or launch vehicles and therefore generally bears the full financial cost of losses. This matters for Starship testing, satellite constellation replacement, launch reliability, government/customer payload commitments, and valuation models that assume smooth scale-up.

21. Legal, Regulatory, And Compliance Review

Legal and regulatory risk is embedded in the core business model. Launch operations require FAA licensing, environmental review, launch-site approvals, airspace coordination, safety compliance, and accident review. Starlink requires FCC and international telecom approvals. Government services require export-control, sanctions, classified-contracting, cybersecurity, procurement, anti-corruption, and national-security compliance. AI/X adds data privacy, AI safety, content moderation, IP, labor, consumer-protection, and platform-governance exposure.

The June 3 S-1/A requires the report to add explicit AI/X regulatory language. The filing references investigations, litigation, and regulatory scrutiny involving Grok and X, including nonconsensual explicit imagery, sexualized content involving women and children, an Irish Data Protection Commission inquiry launched in February 2026 regarding EU personal data including children and Grok within X, FTC chatbot safety inquiries, app-store rules, IP claims, and content moderation obligations. Investors should not treat SpaceX as a clean aerospace/telecom regulatory story.

The litigation section should include Pampena v. Musk. On April 3, 2026, the court entered partial judgment against Musk personally following a March 20, 2026 jury verdict involving two May 2022 Twitter statements. Post-trial motions were filed on May 1, 2026, with a hearing expected in June 2026. The matter is not a SpaceX operating-case driver by itself, but it is relevant to founder-risk, disclosure, reputational, and governance assessment.

The shareholder-rights section should be updated for the Texas Business Court forum-selection provision, venue and procedure restrictions, certain mandatory arbitration provisions, and jury-waiver language. These provisions may reduce litigation uncertainty for the company but can also reduce practical minority-shareholder recourse, which is especially important given controlled-company status and related-party complexity.

Direct-to-device and spectrum risk remains a separate regulatory gate. Starlink Mobile execution depends on spectrum rights, EchoStar spectrum assets, mobile network operator arrangements, handset ecosystem support, and market-by-market regulatory approvals. The business can be valuable, but regulatory access and partner economics can become gating constraints.

Cyber/classified-data exposure should be addressed as a legal and compliance issue, not only as an operational risk. Unauthorized access or disruption could compromise classified or government-controlled information and affect government programs, customer trust, export-control posture, and national-security contracting.

22. Related-Party Transactions And Conflicts

The June 3 S-1/A makes related-party and conflict oversight more important because SpaceX may raise $74.4 billion of base net proceeds while remaining founder-controlled and relying on controlled-company exemptions. A large cash balance can be value-creating if allocated to Starlink, Starship, direct-to-device, and contracted AI infrastructure with high returns. It can be value-dilutive if allocated through opaque related-party structures, speculative long-dated projects, or transactions that primarily benefit the broader Musk ecosystem.

Related-party risk is central to the IPO. The xAI/X merger closed on February 2, 2026, through a 2-step structure that brought X.AI Holdings Corp. into the SpaceX corporate group. The significant subsidiaries list includes X Corp. and X.AI LLC. Given Elon Musk’s central role in SpaceX, xAI, X, Tesla, and related entities, the combination requires careful review of corporate opportunities, intercompany pricing, intellectual property ownership, personnel allocation, compute contracts, shared infrastructure, and customer relationships.

The investor rights agreement identifies Elon Musk and the Elon Musk Revocable Trust dated July 22, 2003, among key holders. The certificate provides founder-specific transfer and control rights. The indemnification agreement excludes certain change-in-control treatment where a transaction is approved by the founder. These provisions reinforce the control architecture and require heightened scrutiny from minority investors.

The founder restricted-stock awards are also conflict-relevant. The awards are large, milestone-based, outside the 2024 Equity Incentive Plan, and carry voting rights from grant date. The milestones include market capitalization, Mars-colony, and non-Earth compute objectives. These awards may align the founder with long-term value creation, but they also increase voting control and create a risk that corporate strategy prioritizes highly ambitious long-dated milestones over near-term capital discipline.

Detailed related-party disclosures now require a Q1 2026 refresh. The S-1/A adds Tesla Megapack purchases, Valor failed-sale-leaseback debt and interest, and a subsequent $6.587 billion Valor AI hardware lease. These are not incidental footnotes because AI capex, power infrastructure, and related-party financing are central to the IPO thesis.

Related-party and strategic-collaboration exposure extends beyond xAI/X integration to Terafab and Cursor. Terafab is described as an initiative with Tesla and Intel, while the Cursor agreement gives SpaceX an option to acquire Cursor at an implied $60.0 billion equity value and creates meaningful termination/deferred-services fee exposure in certain scenarios. Even if strategically coherent, these arrangements increase the need for disclosure on pricing, governance approval, conflict controls, and minority-shareholder protections.

The final report should explicitly map the Musk-ecosystem cash flows because related-party activity is not a footnote at this issuer. The S-1 discloses commercial, services, advertising, lease, aircraft, investment, and strategic relationships among SpaceX, Tesla, xAI, X Holdings, The Boring Company, and Musk-affiliated entities. The issue is not whether these relationships are strategically coherent; it is whether minority shareholders receive enough pricing, approval, and governance evidence to underwrite them.

23. Underwriting, Lock-Up, And Trading Setup

The underwriting and trading setup should be rewritten around the June 3 S-1/A. The expected IPO price is $135.00 per share, the base offering is 555,555,555 primary Class A shares, estimated base net proceeds are $74.4 billion, and underwriters have a 30-day option to purchase 83,333,333 additional Class A shares. The filing says no discount or commission will be paid with respect to shares issued upon exercise of the option. The formal underwriting table still leaves certain final price/discount/proceeds rows blank, so the report should distinguish headline expected terms from final effective pricing.

Lock-up consent is controlled by Goldman Sachs & Co. LLC under the underwriting agreement; Goldman Sachs & Co. LLC may release securities at any time subject to applicable requirements, but Musk shares are not subject to early release and remain locked for 366 days.

Retail and online distribution are now explicit. The filing says representatives may allocate shares to underwriters for sale to their online brokerage account holders and that underwriters do not intend sales to discretionary accounts to exceed 5% of the total Class A shares offered by them. It also anticipates share availability through Charles Schwab & Co., Fidelity Brokerage Services/Fidelity Capital Markets, Robinhood Financial, SoFi Securities, and E*TRADE by Morgan Stanley. These platforms are not affiliated with SpaceX, and platform/app information is not part of the prospectus.

The directed share program should be added to float and supply analysis. Up to 5% of the Class A shares offered may be reserved for selected employees, business relationships, friends and family of executives, and other persons designated at executive-officer discretion. At the base 555.6 million-share offering, that is approximately 27.8 million shares. DSP shares purchased are not subject to lock-up, so they should be treated as potentially immediate float rather than insider-locked supply.

24. Red Flags And Quality Of Earnings Concerns

The most significant red flag is the divergence between Starlink profitability and AI losses. Starlink reportedly generated $4.423 billion of operating profit in 2025, while AI reportedly generated a $6.355 billion operating loss. This suggests that the highest-quality operating asset is funding a large and unproven AI build-out. That may be strategically rational, but it materially changes the equity story.

The second red flag is capex intensity. Reported 2025 capex of $20.737 billion exceeded reported annual revenue of $18.674 billion. AI accounted for 61% of capex. A company can justify this if projects produce contracted, high-return cash flows. Without segment capex, utilization, and contract-quality disclosure, the risk of overinvestment is significant.

The third red flag is Starlink ARPU decline. ARPU declined from $99 in 2023 to $66 in Q1 2026. This does not negate the Starlink growth story, but it requires explanation. Subscriber growth without ARPU stability or contribution-margin data is insufficient for a clean valuation.

The fourth red flag is the $20.0 billion bridge loan and IPO prepayment requirement. The IPO may be partly a balance-sheet refinancing event. Public investors should not assume primary proceeds are entirely available for growth.

The fifth red flag is governance. Class B holders elect 51% of the board while Class B shares remain outstanding. Musk’s 1.302 billion restricted Class B shares disclosed in the beneficial ownership table may be voted by him before vesting, including the January 2026 and March 2026 performance-based awards. Public Class A shareholders will have limited influence.

The sixth red flag is disclosure quality. Customer concentration, retention, churn, RPO, segment gross margin, free cash flow, capex allocation, depreciation policies, AI utilization, and useful-life assumptions are not sufficiently disclosed. These are core valuation inputs.

The seventh red flag is recent structural change. The filing references a common-control reorganization, stock split, and segment reporting change shortly before the IPO filing. This increases comparability risk.

The eighth red flag is the use of an extremely large TAM. A reported $28.5 trillion TAM tied substantially to AI risks inflating the narrative. TAM does not equal monetizable, high-margin, capital-efficient demand.

The most important quality-of-earnings issue is the gap between positive operating cash flow and very large capital expenditures. Operating cash flow was $6.785 billion in 2025, but PPE purchases were $20.737 billion. In Q1 2026, operating cash flow was $1.047 billion and PPE purchases were $10.107 billion. This does not negate the strength of Connectivity, but it does mean consolidated earnings quality must be judged after capex, replacement cycles, financing flows, customer prepayments, and segment-level capital allocation.

25. Bull Case

The strongest bull case is that SpaceX represents a rare global infrastructure platform with durable leadership across launch, satellite broadband, direct-to-device connectivity, government space services, and AI compute. The company has achieved scale in markets where technical failure is common and barriers to entry are high. It is not an early-stage concept stock. It has launched frequently, delivered large payload volumes to orbit, scaled Starlink beyond 10 million subscribers, and built a major AI compute footprint.

Starlink is the core bull-case asset. It has recurring revenue, global reach, rapidly growing subscribers, and reported high operating profit. If ARPU stabilizes and contribution margin remains attractive, Starlink could become 1 of the largest and most profitable global communications infrastructure businesses. The service can serve residential broadband, enterprise mobility, maritime, aviation, government, emergency, and direct-to-device use cases. Its addressable opportunity is large even if the stated TAM is discounted materially.

Launch integration strengthens the bull case. SpaceX’s ability to launch its own satellites creates a cost and cadence advantage. Competitors that must buy launch capacity face higher costs, lower control, and slower deployment. The launch-Starlink flywheel is strategically powerful. Starlink creates internal launch demand, internal demand increases launch cadence, cadence reduces cost and improves reliability, and lower cost enables more constellation capacity.

Starship is a major upside option. If Starship becomes reliable, reusable, and cost-effective, it could materially reduce cost to orbit, increase payload capacity, enable larger satellite deployments, expand government and commercial missions, and support longer-term orbital infrastructure. Starship success could improve Starlink economics and create new revenue categories.

AI could become a large incremental profit pool if SpaceX secures durable customers, high utilization, favorable power economics, and efficient infrastructure. A large compute arrangement with Anthropic would materially increase revenue if the contract is utilized and profitable. The integration of AI, data, platform assets, and infrastructure could create differentiated capability if executed well.

The public-market bull case also includes scarcity value. A company with SpaceX’s strategic importance, technical achievements, founder profile, operating scale, and retail/institutional demand may trade at a premium to conventional valuation frameworks. That premium may persist if growth remains strong and disclosure improves.

The enhanced bull case is that SpaceX converts three currently separate evidence sets into a single compounding infrastructure platform: Connectivity remains highly profitable despite ARPU compression, Starship improves deployment economics for broadband and mobile constellations, and AI infrastructure becomes contracted and utilized at attractive returns. Starlink Mobile adds a real incremental data point to that bull case because it shows early direct-to-device scale rather than only a theoretical addressable market.

26. Bear Case

The strongest bear case is that the IPO valuation already capitalizes too many uncertain outcomes. At a expected $135.00 IPO valuation of approximately $1.75 trillion, the market would be paying for Starlink dominance, launch dominance, Starship success, AI profitability, direct-to-device scale, and long-dated orbital infrastructure optionality. The current disclosed financial profile does not yet prove that those outcomes will generate attractive free cash flow.

The bear case begins with valuation. Approximately 93.7x 2025 revenue is an extraordinary multiple for a company with negative consolidated earnings, capex above revenue, and heavy infrastructure requirements. Public-market investors may initially accept scarcity valuation, but sustained performance will require financial proof.

The second bear-case pillar is AI overinvestment. AI reportedly generated a $6.355 billion operating loss and consumed 61% of consolidated capex in 2025. AI infrastructure markets are competitive and capital-intensive. If customer contracts are cancellable, low-margin, or concentrated, SpaceX may have built expensive capacity without durable cash flow. GPU obsolescence and power constraints can further pressure returns.

The third bear-case pillar is Starlink ARPU compression. Starlink may continue adding subscribers while economics deteriorate. Emerging-market growth, direct-to-device wholesale, promotional pricing, and competition could all pressure ARPU. If margin declines offset volume growth, the Starlink profit engine may not support the consolidated valuation.

The fourth bear-case pillar is governance. Public shareholders have limited influence over capital allocation. Founder control may produce exceptional execution, but it also allows strategic investments that may not maximize near-term public shareholder value. Related-party complexity with xAI, X, Tesla, and other Musk-associated entities increases this risk.

The fifth bear-case pillar is liquidity and dilution. The company has substantial cash, but capex requirements are enormous. If AI and Starship investments continue at scale, additional debt or equity may be needed. The fully diluted share count is complex, and founder awards, employee equity, xAI exchange securities, preferred conversion, warrants, RSUs, SAUs, and ESPP shares can dilute public investors.

The sixth bear-case pillar is regulatory execution. Starship, Starlink, direct-to-device, launch cadence, and international broadband deployment all depend on regulatory approvals. Delays or adverse conditions could slow growth and increase capital costs.

The enhanced bear case is that the profitable Connectivity engine is funding an expanding portfolio of high-capex, low-visibility options while public investors receive weak governance rights. The AI segment's 2025 operating loss, Q1 2026 capex, cancellable Anthropic economics, advertising-heavy revenue mix, and long-dated Terafab/orbital AI claims raise the risk that the consolidated valuation capitalizes unproven option value too aggressively.

27. Base Case And Investment View

The base case is that SpaceX will be received as a scarce mega-cap strategic growth asset rather than a conventionally valued aerospace, telecom, or AI infrastructure company. The June 3 S-1/A gives the market concrete expected IPO economics: $135.00 per share, 555.6 million primary Class A shares, $74.4 billion of base net proceeds, and roughly $1.765 trillion of basic post-offering equity value.

The fundamental base case is more balanced. Starlink/Connectivity is a high-quality asset with real profitability and strong growth. Launch is technically advantaged and strategically reinforcing. Starship provides significant option value. AI is potentially valuable but appears to be the largest source of losses, capex intensity, power demand, and legal/regulatory uncertainty. Governance and dilution remain central because public investors receive limited voting power and immediate dilution of $127.15 per share.

The report should therefore describe the IPO as a rare but expensive asset with a mixed public-equity setup. It is not an obvious avoidance because the company’s assets are real, scaled, and strategically important. It is not an obvious long at the expected $135.00 price because the valuation already capitalizes years of execution across Starlink, Starship, AI compute, direct-to-device, and long-dated non-Earth optionality.

Participation should be contingent on valuation discipline and incremental disclosure. The strongest diligence asks are segment free cash flow, Starlink ARPU/margin stabilization, AI contract duration and customer-credit support, normalized power/cooling economics, maintenance versus growth capex, fully diluted share count, use-of-proceeds waterfall, and independent governance around related-party and founder-linked transactions.

28. What Would Change The View

The view would improve materially if management disclosed audited or reviewed segment-level revenue, gross margin, operating income, depreciation, capex, operating cash flow, and free cash flow across Space, Connectivity, AI, Starlink Mobile, and government services. Segment free cash flow is the single most important missing disclosure.

The view would improve if Starlink ARPU stabilized or if management demonstrated that ARPU decline is mix-driven while contribution margin continues to expand. Cohort-level Starlink retention, churn, gross margin, terminal cost, and payback data would materially increase confidence.

The view would improve if AI customer contracts showed long duration, minimum purchase commitments, limited termination rights, high utilization, favorable power-cost pass-throughs, customer-credit support, and attractive cash gross margins. A large cancellable contract is not sufficient.

The view would improve if the final IPO prospectus confirms $135.00 pricing with a rational valuation discount or if the order book clears at economics that compensate for dilution, governance, and capital intensity. At the current expected price, public investors need more disclosure to justify paying roughly 94.5x 2025 revenue.

The view would worsen if final pricing moves above the expected $135.00 level without incremental disclosure, if IPO proceeds are primarily used for refinancing and opaque related-party commitments, if AI capex continues to outrun durable revenue, if Starlink ARPU declines without margin offset, if legal/regulatory AI/X matters worsen, or if staged lock-up/Form S-8/Rule 144 supply pressures overwhelm initial scarcity demand.

29. Questions For Management

• 1. Confirm whether $135.00 is the final effective IPO price or remains the expected preliminary price, and provide the final underwriting discounts, commissions, and net-proceeds table.
• 2. Provide a full proceeds waterfall across bridge-loan repayment, AI compute, launch infrastructure, satellite constellation scale, spectrum/direct-to-device, working capital, M&A, and general corporate purposes.
• 3. Reconcile March 31 offering-share denominators with May 1 beneficial-ownership denominators, including Class A, Class B, preferred conversion, option exercise, and underwriter option cases.
• 4. Provide the fully diluted share count including options, RSUs, SAUs, warrants, ESPP reserve, founder restricted shares, Form S-8 shares, and shares issuable on Class B conversion.
• 5. Disclose 2023, 2024, 2025, and Q1 2026 segment revenue, gross margin, operating income, depreciation, capex, operating cash flow, and free cash flow.
• 6. Explain AI contract duration, termination rights, minimum commitments, customer concentration, customer-credit support, utilization, power cost, water/cooling cost, and cash gross margin.
• 7. Quantify expected public float by source: base offering, option shares, DSP shares, online brokerage shares, locked shares, 180-day releases, 366-day releases, registration rights, Form S-8, and Rule 144.
• 8. Provide board and committee process materials for related-party transactions, founder awards, AI/xAI/X consolidation, Valor leases, Tesla purchases, and major capital-allocation decisions.
• 9. Provide legal and regulatory exposure summaries for Grok/X, Irish DPC, FTC, app-store rules, nonconsensual explicit imagery, IP claims, Pampena v. Musk, Texas Business Court forum provisions, arbitration, and jury waiver.
• 10. Provide insurance, self-insurance, reserve, and loss-history data for launch vehicles, payloads, satellites, Starship tests, and constellation replacement.

30. Data Room And Follow-Up Diligence Requests

The data room should include audited historical segment financials for 2023, 2024, 2025, and Q1 2026; pro forma financial statements for SpaceX plus xAI/X; a reconciliation from standalone SpaceX to the combined entity; and segment-level revenue bridges by price, volume, mix, geography, customer type, and acquisition contribution.

IPO-specific diligence should include the final effective prospectus, final underwriting price/discount/proceeds table, full proceeds waterfall, bridge-loan repayment schedule, pro forma capitalization by class, base and full-option share counts, full diluted share-count schedule, new-investor ownership/consideration bridge, and net tangible book value calculation.

Trading and supply diligence should include DSP allocation, online brokerage allocation, discretionary-account sales, underwriter option exercise assumptions, detailed lock-up shareholder schedules, Form S-8 registration scope, registration-rights agreements, and Rule 144 eligibility calendars.

AI diligence should include customer contracts, minimum commitments, termination rights, utilization schedules, power contracts, GPU procurement contracts, depreciation schedules, networking costs, data-center capex, customer concentration, service-level obligations, gross margin by cluster, legal/regulatory exposure, and customer-credit support.

Governance diligence should include board minutes, special-committee records, fairness materials, related-party transaction approvals, compensation committee materials, founder-award approval documentation, corporate opportunity policies, Texas forum/arbitration/jury-waiver analysis, and independent director analysis.

Regulatory and risk diligence should include FAA correspondence, FCC correspondence, international telecom approvals, export-control compliance records, environmental review materials, national-security contracting compliance, cybersecurity audits, AI safety audits, privacy/data-governance policies, insurance/self-insurance data, and launch/satellite loss history.

31. Final Investment Committee Takeaway

SpaceX is an extraordinary asset but not a simple IPO. The June 3 S-1/A makes the transaction far more concrete: 555,555,555 primary Class A shares, an expected $135.00 price, $74.4 billion of estimated base net proceeds, an 83,333,333-share underwriter option, approximately $1.765 trillion of basic post-offering equity value, $7.85 of pro forma as-adjusted NTA/share, and $127.15 of immediate dilution to new investors.

The company deserves immediate deep diligence. Avoidance would be premature given the quality, scale, and strategic importance of the assets. Unconditional participation would also be premature at the expected price. The correct posture is valuation-sensitive engagement, with participation contingent on segment-level free-cash-flow proof, durable Starlink economics, credible AI margins and contract duration, disciplined proceeds allocation, and governance/risk terms that compensate public investors for limited voting rights.

The updated S-1/A-based view is more segmented and more demanding. Connectivity is real, profitable, and scaled. Space and Starship are strategically essential but capital-intensive. AI is potentially valuable but currently mixes advertising, platform distribution, compute infrastructure, heavy capex, cancellable contract exposure, power/turbine constraints, regulatory scrutiny, and legal risk. The $135.00 IPO price requires the market to capitalize much of that future value upfront.

The report should therefore be updated across text, tables, and date references to make the June 3 S-1/A the controlling filing. The June 1 S-1/A should be retained only as historical context; the current investment, valuation, dilution, underwriting, lock-up, and proceeds analysis should reference the Form S-1/A filed June 3, 2026.