SpaceX has been the most important company in aerospace for a decade, and until this week, no one outside its boardroom knew its actual profit margins. The IPO roadshow that begins today changes that, releasing a set of financial disclosures that would have been unimaginable from Elon Musk even three years ago. The company is targeting a $1.77 trillion valuation. Morningstar responded by publishing a note saying SpaceX is worth approximately half that.
What Actually Happened
On June 4, 2026, SpaceX released a 17-minute IPO roadshow video presented by CFO Brett Johnson, alongside an S-1/A filing with the Securities and Exchange Commission containing the most detailed financial disclosures in the company's 24-year history. The filing revealed that Starlink generated $11.4 billion in revenue in 2025, representing 61% of SpaceX's total company revenue. In Q1 2026, Starlink's revenue share rose to 69%, with total company revenue for the quarter at approximately $5.3 billion. The company plans to sell 556.6 million shares at a target price of $135 per share, implying a total fundraise of $75 billion and a valuation of $1.77 trillion at the offer price. Pricing is scheduled for June 11, with SPCX shares beginning trading on Nasdaq on June 12.
The margin targets disclosed in the roadshow video are the most consequential numbers in the filing and have received less coverage than the topline valuation figures. SpaceX disclosed that it is targeting gross margins of approximately 70%, up from 49% currently, and a long-term net profit margin target of 45%. These numbers are exceptional for any capital-intensive hardware company. Boeing's gross margins typically run between 5% and 10%. Airbus operates at roughly 8% to 12%. SpaceX's 70% gross margin target is closer to enterprise software economics than to aerospace manufacturing, driven primarily by Starlink's near-zero marginal cost of serving an additional subscriber once the satellite constellation is deployed.
The subscriber data disclosed alongside the financials confirms the scale of what Starlink has become. As of March 31, 2026, Starlink had 10.3 million subscribers across 155 countries. Average revenue per user (ARPU) has increased consistently over the past eight quarters as SpaceX launched premium service tiers for maritime, aviation, and government customers. The premium segments now account for over 15% of Starlink service revenue at approximately three times the residential ARPU. The 10.3 million subscriber base compares to Amazon Project Kuiper's pre-commercial status and Eutelsat OneWeb's roughly 650 enterprise customers, confirming Starlink's structural lead in satellite broadband at consumer scale.
Why This Matters More Than People Think
SpaceX's IPO is not primarily an aerospace story. It is the first public market test of whether AI infrastructure warrants software-like valuations on hardware economics. Starlink is the compute backbone for remote AI operations, for government intelligence infrastructure in regions without terrestrial connectivity, and for the consumer broadband segment in areas where fiber deployment is not economically viable. Google's $920 million monthly compute partnership with SpaceX, disclosed in May 2026, confirmed that Starlink is now categorized as AI infrastructure by the market's most sophisticated buyers, not as consumer internet. That reclassification is what justifies the 70% gross margin target in investor conversations: infrastructure businesses with low marginal costs and high switching costs trade at software multiples, not aerospace multiples.
Morningstar analyst David Whiston published a valuation note on June 3, 2026, calling the $1.75 trillion IPO target "nearly twice fair value." His discounted cash flow model values SpaceX at $780 billion, approximately 48% below the offer price. His methodology applies standard aerospace and satellite communications industry multiples to Starlink's current revenue trajectory and the Falcon 9 launch manifest. He explicitly excludes Starship's commercial potential from the base case, which is the methodologically conservative choice for a DCF model: you do not credit unproven cash flows. The $780 billion estimate represents what SpaceX is worth if Starship never generates commercial revenue at scale and Starlink grows at its current pace without competitive disruption.
The IPO creates a public valuation anchor for the entire private AI infrastructure market, and that is its most consequential function beyond the fundraise itself. If SPCX opens trading on June 12 at a premium to its $135 offer price, which institutional demand indicators currently suggest, it validates the $300 billion to $500 billion private valuations being assigned to companies like Cerebras (which IPO'd in May 2026 at a $95 billion peak market cap), Anduril, and Shield AI. If SPCX opens below the offer price, those private valuations are immediately subject to markdown pressure, because they rest on the same fundamental assumption: that infrastructure companies enabling the AI era deserve software-like valuations on hardware economics.
The Competitive Landscape
Starlink faces its most credible competitive challenge yet, arriving within months of its public market debut. Amazon's Project Kuiper achieved its 1,000th satellite milestone in May 2026 and is targeting commercial service launch in Q3 2026 across North American and European markets. OneWeb, integrated into the Eutelsat SoftBank constellation, has 648 operational satellites covering 90% of global landmass for enterprise customers. The competitive differentiation is unit economics: SpaceX manufactures both its launch vehicles and its Starlink satellites internally, giving it an estimated 60% to 70% cost advantage per deployed satellite versus any competitor that purchases external launch services. That cost advantage compounds as Starship replaces Falcon 9 for constellation replenishment missions.
The financial disclosures reveal a business architecture that existing satellite companies have never achieved. SpaceX's rocket launch business, which generates revenue from commercial, government, and internal Starlink deployment missions, effectively cross-subsidizes Starlink's subscriber acquisition and constellation expansion. Every Falcon 9 mission flown for a paying external customer offsets a portion of the capital cost of deploying additional Starlink satellites. This internal cross-subsidy is why SpaceX can price residential Starlink service at $120 per month in competitive markets while simultaneously investing in a 40,000-satellite constellation upgrade. No competitor has this structural advantage. Amazon has purchasing power and distribution scale, but it does not manufacture its own rockets at the volumes required to match SpaceX's internal launch economics.
The Amazon AWS parallel from 2015 is the most instructive historical comparison available. When Amazon disclosed AWS financials for the first time in Q1 2015, it revealed a business generating 25% operating margins inside a company that the market had been valuing as a near-zero-margin retailer. The disclosure changed how investors valued Amazon's entire enterprise: it revealed that the retail business was effectively being cross-subsidized by a cloud services business with software economics. SpaceX's Starlink disclosure does something structurally similar: it reveals that the rocket launch business is being cross-subsidized by a software-margin satellite internet business. That business model architecture, hardware at the infrastructure layer enabling software economics at the service layer, has never been tested in public aerospace markets before this IPO.
Hidden Insight: The Starship Variable Morningstar Cannot Model
Morningstar's $780 billion fair value estimate explicitly excludes Starship's commercial potential from its base case model. This is the methodologically correct decision for a DCF analysis: discounted cash flow models do not credit speculative future cash flows without contractual support. But the exclusion creates a valuation floor that misses what could be the most transformative capability shift in modern transportation economics. Starship achieved its first successful commercial payload delivery in Q1 2026, and the current manifest for 2027 includes 14 booked commercial launches at an average contract price of $85 million per mission. That is $1.19 billion in booked 2027 Starship revenue that is visible today and is not reflected in Morningstar's model.
The deeper insight is about what Starship enables for the AI industry's infrastructure trajectory specifically. SpaceX's internal models project that Starship reduces the cost-per-kilogram to low Earth orbit from Falcon 9's approximately $2,700 per kilogram to under $100 per kilogram at full reusability and high launch cadence. At that cost structure, deploying a next-generation Starlink constellation from its current 6,800 satellites to the 40,000-plus satellite configuration in SpaceX's FCC license becomes economically accessible on a timeline of three to five years rather than a decade. A 40,000-satellite constellation at current Starlink orbital altitudes creates an always-on, low-latency global network with aggregate bandwidth that rivals terrestrial fiber infrastructure in major metropolitan areas, which is the compute backbone that AI inference at the edge requires at scale.
The AI-at-the-edge thesis is where SpaceX's valuation premium lives and where Morningstar's model has a structural blind spot. Today's AI models run primarily in centralized data centers operated by hyperscalers. The competitive frontier for 2028 to 2030 is inference at the point of need: autonomous vehicles requiring sub-50ms latency without cellular infrastructure, field robotics in remote industrial environments, defense applications that cannot depend on terrestrial communications that can be jammed or destroyed, and consumer AI assistants that function when fiber and cellular networks are unavailable. SpaceX is the only company currently positioned to provide that infrastructure globally at the bandwidth, latency, and cost structure required for these use cases. That optionality has a real expected value even if it resists precise modeling.
The bear case, however, is exactly what Morningstar is arguing: optionality does not pay dividends, and SpaceX's current trajectory does not justify a $1.75 trillion entry price on any conventional valuation framework. The company's actual 2025 revenue was approximately $18.7 billion with an estimated net margin of roughly 10%, implying a price-to-earnings multiple of over 100x at the IPO price. For that multiple to be justified, SpaceX needs to execute on the Starlink subscriber ramp to 15 million subscribers by year-end 2026, the Starship commercial launch program at scale, and the margin expansion from 49% to 70% gross simultaneously. Skeptics point out that executing any two of those three on the required timeline would be an exceptional execution result for any company; executing all three concurrently requires sustained operational excellence that SpaceX has demonstrated historically but never at this financial scale or under public market scrutiny.
What to Watch Next
June 11 is the IPO pricing date and June 12 is the first trading day on Nasdaq under the ticker SPCX. The first-day price action will resolve the most important near-term question: does retail demand, which is expected to be enormous given Musk's retail investor following, override institutional skepticism about the valuation gap? A day-one trade above $150 signals that the Morningstar bear case is being treated as a minority view by the market. A flat open around $135 signals that institutional investors are not bidding it up, which would be unusual for a hyped IPO but would validate Morningstar's concern that the offer price is at or near fair value even on optimistic assumptions.
The 90-day indicator is Q2 2026 subscriber and revenue disclosure, which SpaceX will report as its first public earnings release in August 2026. The internal model underlying the IPO assumes Starlink reaches 12 million active subscribers by year-end 2026, implying Q2 needs to come in at approximately 11 million active subscribers to stay on pace. If Q2 comes in below 10.5 million active subscribers, the growth assumptions embedded in the $1.75 trillion valuation are immediately questioned by institutional shareholders and the stock faces selling pressure before the first anniversary of its listing.
The 180-day question is whether Amazon Kuiper's Q3 2026 commercial launch materially affects Starlink pricing dynamics. SpaceX's residential service currently runs at $120 per month in most developed markets. Amazon has the balance sheet to subsidize Kuiper customer acquisition aggressively and the distribution infrastructure through Prime to offer bundled broadband and retail services that SpaceX cannot match. If Kuiper launches at a competitive price point below $100 per month, it tests whether Starlink's pricing power is structural (manufacturing cost advantage and established subscriber base) or primarily the result of being the only viable satellite broadband option in most markets. That competitive test will be the most important datapoint for SPCX's valuation trajectory through year-end 2026 and into early 2027.
SpaceX just told the world what it is really worth. The world believes it. Morningstar does not. The market opens trading on June 12 and one of them is right.
Key Takeaways
- Starlink generated $11.4 billion revenue in 2025: Representing 61% of SpaceX's total revenue and rising to 69% in Q1 2026; these are the first detailed financials SpaceX has ever disclosed publicly.
- SpaceX targets 70% gross margins versus 49% today: The long-term 45% net profit margin target places SpaceX's financial model closer to enterprise software economics than to any aerospace comparables.
- Morningstar values SpaceX at $780 billion, 48% below the offer price: The model excludes Starship commercial potential; the gap between $780B and $1.75T is where the market debate about AI infrastructure valuations will be resolved.
- 10.3 million Starlink subscribers across 155 countries: The subscriber base is growing with premium maritime and aviation tiers now representing over 15% of service revenue at 3x the residential ARPU.
- IPO creates a public valuation anchor for private AI infrastructure: If SPCX trades at a premium to its $135 offer price, it validates $300 billion to $500 billion private valuations for companies like Cerebras and Anduril; a flat open puts those valuations under immediate pressure.
Questions Worth Asking
- If Starship reduces launch costs to under $100 per kilogram, does the entire satellite communications industry become economically non-viable for any company that does not manufacture its own launch vehicles?
- Does the public market's willingness to pay 100x earnings for SpaceX signal that investors have fundamentally reframed satellite infrastructure companies as software businesses, and what happens to that reframing if margin expansion from 49% to 70% takes longer than projected?
- Amazon Kuiper's Q3 2026 launch will test whether Starlink's pricing power is structural or opportunistic: what evidence should investors watch for in the 90 days after Kuiper's commercial debut to determine which interpretation is correct?
