SpaceX

June 11, 2026 — the night before the Nasdaq debut

There is a pattern that repeats itself in every era of transformative technology. A company arrives that is genuinely extraordinary — one that has done things no competitor has done, controls assets no competitor controls, and operates businesses that are already changing the world. The market, correctly recognizing the extraordinary nature of the company, then makes a second, separate, and far more questionable judgment: that extraordinary companies justify any price. The dot-com era gave us this lesson with Cisco, which was genuinely building the internet's backbone and genuinely deserved a premium valuation, but was priced at 130x earnings in March 2000 in a way that took thirteen years to recover from. The lesson was not that Cisco was a bad company. It was that even great companies can be terrible investments at the wrong price.

SpaceX is the Cisco of 2026 — except the ambition is larger, the optionality is more genuine, and the price is more extreme.

Tonight, SpaceX prices its IPO at a fixed $135 per share, raising $75 billion in what is officially the largest IPO in global corporate history, surpassing Saudi Aramco's $29.4 billion raise in 2019. The implied post-offering equity value is approximately $1.77 trillion. The order book closed 3.5x to 4x oversubscribed, with more than $250 billion in total demand. Retail investors alone submitted over $100 billion in orders. BlackRock placed an anchor order of at least $5 billion. Tomorrow morning, SPCX begins trading on the Nasdaq.

My job is not to tell you whether to buy or sell this stock. My job is to tell you what the numbers say about whether $135 is justified by intrinsic value — and what you would have to believe to disagree with that assessment. The answer, stated plainly, is that $135 is not justified by intrinsic value under any reasonable probability-weighted DCF framework. It is justified only if you assign near-certainty to outcomes that are today closer to engineering ambitions than demonstrated cash flows.

1.The Three-Business Problem

Before any valuation, the analyst must understand what is actually being purchased. SpaceX is not one company. It is three companies at radically different stages of maturity, profitability, and capital intensity, consolidated under one S-1 through a common-control merger with xAI and X Corp finalized on February 2, 2026.

The segment financials from the S-1 are unambiguous:

Notice how the segment picture destroys the simple narrative. The company that the IPO marketing materials present as a unified AI-launch-connectivity juggernaut is, in its current financials, a company where one business (Starlink) generates $4.4 billion in operating income, a second business (launch) is temporarily loss-making due to Starship R&D, and a third business (AI/xAI) loses $6.4 billion on $3.2 billion in revenue while consuming $12.7 billion in capital expenditure. Total capex of $20.7 billion exceeded consolidated revenue of $18.7 billion in FY2025. This is not a criticism of the strategy — it may be the right strategy — but it is a fact that the valuation must confront.

2.The Story Behind the Numbers

The table above maps the qualitative narrative to the specific numerical inputs that drive the DCF. The central story for SpaceX is this: Starlink is the anchor of current intrinsic value, launch is a temporarily depressed but structurally dominant franchise, and AI/xAI is the enormous swing factor whose ultimate economics are genuinely unknowable today.

R&D capitalization is the first important modeling choice. Starship development, satellite manufacturing improvements, Starlink network upgrades, and Grok model development are all intended to create multi-year revenue capacity rather than only current-period benefits. Expensing them as incurred, as GAAP requires, distorts the true economics. The S-1 shows $8.6 billion in R&D expense in FY2025, of which $3.0 billion was Starship-related within the Space segment and $5.1 billion was AI/xAI-related. Capitalizing these amounts and amortizing them over their useful lives is not aggressive accounting — it is the economically correct treatment, and it is what I have done throughout this analysis.

3.Building the Valuation: Segment by Segment

Space / Launch

The launch business is the simplest to understand and the hardest to value because Starship sits directly in the middle of both its near-term losses and its long-term potential.

SpaceX executed approximately 170 missions in FY2025, representing roughly 51% of all global launches and over 90% of Western commercial market share. Mass to orbit reached 2,213 metric tons in 2025, up from 1,699 in 2024. The Falcon 9 reusability advantage is real and already demonstrated. The cost structure is strong: cost of revenue fell to 33.1% of Space segment sales in FY2025, implying gross economics that would be very attractive if not for the $3.0 billion in Starship R&D being expensed through the segment.

The year-10 launch revenue target of $35 billion assumes SpaceX remains the dominant Western launch provider and Starship expands the addressable market beyond Falcon-style missions, but captures less than 10% of the S-1's $370 billion space/launch TAM. That is a conservative assumption given current market share, but it reflects real limits: cadence, regulatory approvals, competitive responses from Rocket Lab's Neutron and Blue Origin, and the practical ceiling on government contract growth.

Target margin rationale: Normalized Falcon/government launch economics at 30% margin × 70% mix, plus Starship heavy-lift and space-infrastructure services at 38% margin × 30% mix, gives a weighted average of 32.4%. Subtracting 0.4 percentage points for insurance, mission-failure provisions, and government-pricing risk gives a 32.0% target operating margin.

Capital intensity: Launch capex was approximately $3.83 billion on $4.09 billion of FY2025 sales — only about 1.1x sales-to-capital — because Starship development is the dominant investment. As Starship commercializes and fixed launch infrastructure is utilized at higher cadence, capital efficiency should improve. Sales-to-capital of 1.4x in years 1–5, improving to 3.0x in years 6–10, reflects this trajectory.

Connectivity / Starlink

Starlink is the best business in this filing. Full stop.

Operating margin expanded from 12.1% in 2023 to 38.8% in 2025, with adjusted EBITDA margins reaching 63%. Subscribers grew from 5.0 million in Q1 2025 to 10.3 million by Q1 2026 — a 105% year-over-year increase. The business operates across more than 160 countries. Satellite manufacturing cost per Gbps has declined approximately 3x, with a target of 9x by Starlink V3. Starlink kit costs have fallen 59% since 2022.

The complication is ARPU. Monthly revenue per user has compressed from $99 in 2023 to $81 in 2025 to $66 in Q1 2026. This compression reflects international expansion into lower-income markets, which is the right strategic move but creates a revenue quality question. A subscriber in rural Nigeria is not the same as a subscriber in suburban Texas.

The year-10 connectivity revenue target of $105 billion assumes Starlink continues scaling broadband, enterprise, maritime/aviation, government, and direct-to-device offerings globally, reaching approximately 6.6% of the S-1's $1.6 trillion connectivity TAM. The ARPU compression is already embedded in that assumption.

Target margin rationale: Residential and business broadband at 42% margin × 65% mix, plus enterprise/government/mobility at 50% margin × 25% mix, plus direct-to-device/wholesale at 38% margin × 10% mix gives a weighted average of 43.6%. Adding 1.4 percentage points for satellite and terminal cost declines, then subtracting 1.0 percentage point for ARPU compression and spectrum/regulatory costs, gives a 44.0% target operating margin.

Capital intensity: FY2025 capex of $4.18 billion on $11.39 billion of sales implies approximately 2.7x current sales-to-capital. As the constellation matures and deployment shifts from build-out to replacement cadence — enabled by Starship's lower cost-per-kg — capital efficiency should improve toward 4.2x in years 6–10. This remains more capital-intensive than terrestrial software because satellites require periodic replacement.

AI / xAI / X

This is where the valuation becomes genuinely difficult, and where the honest analyst must resist the temptation to extrapolate from the most exciting scenario.

The AI segment today is a capital destruction engine. In FY2025, it generated $3.2 billion in revenue against $6.4 billion in operating losses and $12.7 billion in capex. In Q1 2026 alone, AI capex reached $7.7 billion against $818 million in revenue. The segment has deployed 1.0 GW of nameplate compute draw across the COLOSSUS and COLOSSUS II facilities. It has 550 million monthly active users on X and 1.9 million SuperGrok paid subscribers.

The Anthropic cloud services agreement, at $1.25 billion per month ($15 billion annualized) for access to 325,000 NVIDIA GPUs, is the most important near-term data point. The Google agreement, at $920 million per month running from October 2026 through June 2029, adds further revenue visibility. Together, these contracts could push the AI segment's annualized revenue run rate toward $25–$30 billion. The critical uncertainties are contract durability — the Anthropic agreement reportedly contains 90-day rolling termination features — and gross margin, which is not disclosed but could be thin if SpaceX is primarily arbitraging scarce GPU capacity rather than operating a differentiated platform.

The year-10 AI revenue target of $180 billion gives SpaceX credit for the Anthropic and Google compute agreements, X distribution, Grok subscriptions, and some AI infrastructure platform growth, but reaches less than 1% of the S-1's $26.5 trillion AI TAM. That TAM is heavily inflated by including the entire global "knowledge work" and digital economy market.

Target margin rationale: Terrestrial AI compute infrastructure at 25% margin × 65% mix, plus Grok/software and enterprise AI at 40% margin × 25% mix, plus early Terafab/orbital-enabled services at 45% margin × 10% mix gives a weighted average of 30.75%. Subtracting 2.75 percentage points for GPU pricing normalization, contract-renewal risk, and intense hyperscaler competition gives a 28.0% target operating margin. This haircut is explicit and necessary: the AI infrastructure market will face OpenAI, Google, Microsoft, AWS, Meta, and Nvidia-ecosystem competitors with deeper enterprise relationships and larger balance sheets.

Capital intensity: AI capex was $12.73 billion on $3.20 billion of FY2025 sales, or only 0.25x sales-to-capital. Contracted compute demand improves this ratio, but data centers, chips, power, Terafab, and potential orbital infrastructure keep this segment the dominant capital sink. Sales-to-capital of 0.9x in years 1–5, improving to 1.6x in years 6–10, reflects the heavy ongoing investment required.

4.Segment Assumptions Summary

Consolidated margin reconciliation: The revenue-weighted average of segment target margins at year 10 — ($35B × 32.0% + $105B × 44.0% + $180B × 28.0%) ÷ $320B = ($11.2B + $46.2B + $50.4B) ÷ $320B = $107.8B ÷ $320B = 33.69% — which matches the consolidated target margin of 33.6875% used in the DCF. Every dollar in the model can be traced back to these segment-level assumptions.

5.The DCF Model

Here is the entire model on one page. The key inputs and their rationale:

Revenue growth: Year-1 growth of 36.5% reflects Starlink subscriber expansion, AI contract ramp, and a partial rebound in launch timing. Years 2–5 growth of 41.4% gives weight to continued Starlink global penetration, Starlink Mobile, Starship-enabled launch demand, and scaling of terrestrial AI compute — while treating Terafab and orbital data centers as partial contributors rather than fully proven cash flows. Years 6–10 growth of 49.4% is driven by AI/xAI becoming the largest segment and Starlink compounding from a much larger base. This path reaches approximately $320 billion of revenue by year 10.

Operating margins: Year-1 margin of -16.6% reflects the current consolidated reality — Starlink's profitability is more than offset by Starship development spending and very large AI/xAI operating losses. Margins converge to 33.7% over 10 years as Starship R&D moderates, Starlink scales, and AI compute utilization improves. The 10-year convergence timeline is deliberately conservative: Starship commercialization, Starlink V3/mobile scaling, and AI data-center utilization all require long engineering and capital-deployment cycles.

Cost of capital: Initial WACC of 7.76%, rising to 8.96% at terminal. The starting rate reflects the mix of contracted government revenue, subscription connectivity, and speculative AI investment. The terminal rate is higher because the mature company will have a less favorable risk profile than the early-growth phase when Starlink's contracted cash flows dominate.

Terminal value: Terminal growth of 4.0% is below the 4.46% risk-free rate, making it sustainable as a perpetual nominal growth assumption. Terminal ROIC of 13.0% exceeds the terminal cost of capital of 8.96%, preserving value creation in perpetuity but not dramatically so. One important caution: terminal value represents 97.3% of total model value. This is high even for a young, scaling company, and it means the valuation is acutely sensitive to assumptions about long-run margins and reinvestment needs.

The cash flow chart illustrates the central challenge: free cash flow to the firm is deeply negative through approximately year 6, as reinvestment consistently exceeds after-tax operating income. The model projects FCFF turning positive around year 7, with meaningful positive cash generation only in years 8–10. This is not a company that will generate significant free cash flow in the near term under any reasonable assumption set.

The DCF result: $44.23 per share. Against an IPO price of $135.00, this implies approximately 67% downside to intrinsic value. The equity value in the model is approximately $554 billion against an IPO-implied equity value of $1.77 trillion.

The complete Excel valuation model, with all segment assumptions, cash flow projections, and sensitivity scenarios, accompanies this report.

6.What the Valuation Bridge Tells Us

The valuation bridge decomposes the gap between my DCF value and the IPO price into its component parts. The $90.77 gap between $44.23 and $135.00 is not simply "market optimism." It represents specific implicit beliefs about AI scale, Starship cadence, Terafab execution, and orbital compute feasibility that the market is pricing as high-probability outcomes.

The bridge also explains why the IPO is 3.5x–4x oversubscribed. Investors are not all making the same bet. Some are buying Starlink's proven economics. Some are buying the launch monopoly. Some are buying AI infrastructure optionality. Some are buying scarcity value in a low-float mega-cap. Some are buying the Musk-platform narrative premium. The $135 price is the sum of all those bets simultaneously — and the question for a long-term investor is whether the sum is coherent.

7.The Terafab and Orbital Compute Question

The thesis explicitly highlights two initiatives that could, if successful, transform the AI segment's unit economics: the Terafab chip-manufacturing collaboration with Tesla and Intel, targeting one terawatt of compute hardware per year, and orbital data centers powered by space-based solar energy.

These are genuinely interesting ideas. The Terafab initiative, announced in March 2026 with Intel joining in April, targets a 100-million-square-foot closed-loop manufacturing campus in Texas. The stated investment plan is up to $122 billion for the full-scale facility, with $55 billion for the initial prototype phase. The orbital AI data center concept, for which SpaceX has filed FCC applications to deploy up to 1 million satellites, exploits the fact that space-based solar arrays capture roughly 5x the energy per unit area of terrestrial solar due to continuous illumination and lack of atmospheric interference, while radiative cooling in vacuum eliminates water consumption and cooling costs.

The engineering logic is sound. The economic logic is compelling in theory. The execution risk is enormous in practice.

On Terafab: semiconductor fabrication is among the most capital-intensive and technically demanding industries in the global economy. Intel's partnership reduces execution risk but does not remove it. Independent semiconductor analysts have estimated full-scale costs far above the $122 billion stated in local tax filings — with some estimates ranging from $5 trillion to $13 trillion for the complete vision. Even at the stated $122 billion, this initiative alone would consume the entire IPO proceeds with nothing left for Starship, satellite deployment, or AI data centers.

On orbital compute: the thermal engineering problem is not trivial. AI workloads produce heat loads of 120–150 kW per satellite, and vacuum offers no convective cooling. The radiative panels required to dissipate that heat would need to be wider than a Boeing 747 per satellite. Radiation hardening, thermal cycling, chip degradation, and on-orbit maintenance are all unresolved at commercial scale. There is no demonstrated operating orbital AI data center.

My treatment of these initiatives in the model is explicit: they are real options, not base-case cash flows. The 28% target margin for the AI segment includes only 10% weight on early Terafab/orbital-enabled services at a 45% margin. The remaining 90% of the AI segment is valued on terrestrial compute and software economics. If Terafab and orbital compute work as described, the upside is substantial. If they encounter the capital overruns and engineering obstacles that characterize most first-of-kind physical infrastructure projects, the base case is not impaired.

The cross-business synergy logic — Starship launch capacity enabling satellite deployment, Starlink laser mesh routing AI workloads to orbital compute, AI optimizing Starlink constellation management — is genuinely compelling as a strategic architecture. SpaceX is the only company that controls all the links in that chain. The question is not whether the chain is theoretically powerful. The question is whether each link will be operational, at scale, within the valuation horizon.

8.The Reverse DCF: What the Market Believes

The most revealing exercise in any IPO valuation is the reverse DCF: rather than asking what the company is worth, ask what the company would have to achieve to justify the price being offered. The table above shows my estimates against the market-implied assumptions at the $135 price.

The reverse DCF converged at $132.01 per share using the following assumptions:

• Years 2–5 revenue growth: 60.0% (versus my base of 41.4%)
• Target operating margin: 45.0% (versus my base of 33.7%)
• Terminal growth rate: 5.75% (versus my base of 4.0%)

Notice that last number carefully. A terminal growth rate of 5.75% is above the current risk-free rate of approximately 4.46%. In a Damodaran-style framework, a company cannot grow faster than the economy in perpetuity — that is a mathematical impossibility. A terminal growth rate above the risk-free rate implies the company eventually becomes larger than the global economy. The market is embedding an assumption that is not just aggressive — it is technically inconsistent with a stable, mature company.

The 45% target operating margin is also striking. My base case assumes 33.7%, which I already consider ambitious for a company with aerospace manufacturing, satellite operations, telecom infrastructure, and AI data centers. The market is implicitly pricing SpaceX as if it will achieve operating margins closer to a pure software platform. That requires the AI segment to achieve margins far above my 28% base case — which in turn requires either Terafab to work, orbital compute to work, or both.

New Constructs has made the same observation from a different angle: justifying the $1.77 trillion valuation at a 10% annual return threshold requires revenue of approximately $1.1 trillion by 2035, implying a 50% CAGR for a decade — a pace never achieved in corporate history, which would make SpaceX alone represent 2.4% of projected U.S. GDP.

9.Sensitivity Analysis: The Full Range of Outcomes

The sensitivity analysis tests the model across the full range of plausible assumptions. The results are unambiguous:

The highest value in the tested sensitivity range is $55.59 per share — still 58.8% below the $135 IPO price. The model is not built from an extreme bearish corner of the assumption set. Even the most optimistic tested scenario does not approach the offer price.

The two dominant variables are medium-term revenue growth and mature operating margin. This makes intuitive sense: the AI segment's scaling trajectory in years 2–5 and the long-run margin profile of the combined business are the two assumptions that are simultaneously most uncertain and most consequential. Governance risk, capital intensity, and terminal growth rate matter, but they are secondary to getting the AI revenue and margin story right.

10.Peer Context: Why Comparables Are Difficult But Instructive

No single public company is comparable to SpaceX. The business combines a launch franchise, a satellite connectivity network, a defense and government contractor, a consumer broadband provider, and an AI infrastructure platform. Mapping any of those segments to public peers illustrates why the IPO multiple is difficult to defend:

SpaceX at $135 trades at roughly 91x trailing enterprise value to sales. Even Nvidia, at its 2024 peak when it was the fastest-growing large-cap in the world, traded at approximately 20–25x sales. The SpaceX multiple is not a peer premium — it is a different category of valuation that requires the AI/orbital-compute optionality to be worth more than all of the proven businesses combined, several times over.

At my DCF fair value of $44.23, SpaceX trades at approximately 30x trailing EV/Sales — still a substantial premium to any peer group, but one that reflects the genuine quality of Starlink and the launch franchise without requiring Terafab or orbital compute to succeed.

11.The Governance Discount That the Market Is Ignoring

One element of the IPO that deserves more attention than it is receiving is the governance structure. Elon Musk retains approximately 82.4% of combined voting power through a dual-class structure where Class B shares carry 10 votes each. SpaceX qualifies as a controlled company and has opted out of Nasdaq's independent board requirements. Musk can only be removed as CEO and Chairman by a majority vote of Class B shares — which he controls.

The xAI/X consolidation into SpaceX is the most acute governance concern. The transaction was accounted for under common-control rules (ASC 805-50), which permitted SpaceX to absorb xAI and X without allocating goodwill or fair-valued intangibles. This means approximately a quarter-trillion dollars of economic premium paid by SpaceX shareholders is invisible on the balance sheet. Morningstar has labeled this a "highly speculative related-party transaction not conducted at arm's length," and Senator Elizabeth Warren formally requested an SEC delay of the IPO citing governance issues and lack of minority investor protection.

The compensation structure disclosed in the S-1 is also remarkable. Musk received 1 billion restricted Class B shares vesting across 15 tranches contingent on market capitalization milestones from $500 billion to $7.5 trillion AND the establishment of a permanent human colony on Mars with at least 1 million inhabitants. He also received 302 million additional restricted Class B shares vesting contingent on market cap milestones AND the completion of non-Earth-based data centers capable of delivering 100 terawatts of compute per year.

These are not standard performance metrics. They are milestones that align Musk's personal wealth with outcomes that may require SpaceX to allocate capital to Mars colonization and orbital compute regardless of whether those investments maximize returns for public shareholders. A governance discount of 100–300 basis points on the cost of equity is not unreasonable given this structure. My model does not apply a separate governance haircut — the risk is partially embedded in the cost of capital — but investors should recognize that the margin of safety for minority shareholders is structurally limited.

12.Analyst Consensus: There Is None, and That Is Telling

Because SpaceX is pricing as a private company on June 11, 2026, no sell-side analyst consensus exists yet for SPCX. The available external reference points are:

The range of serious fundamental estimates runs from $44 to $100. The IPO price sits above all of them. The bullish analyst targets of $165–$195 rely on the orbital compute and Terafab scenarios materializing within a 12–18 month horizon — an assumption that the engineering timelines and capital requirements do not support.

The absence of a sell-side consensus is itself informative. When a company prices at $1.77 trillion with no established public earnings history, no analyst coverage, and no price discovery through a traditional book-building process, the market is operating on narrative rather than consensus fundamental analysis. That is not necessarily wrong — narratives can be correct — but it means the price has not been stress-tested by the usual mechanisms.

13.The Final Valuation Summary

The DCF model values SpaceX at $44.23 per share. The IPO prices it at $135.00. The difference is not primarily about discount rates or terminal growth assumptions — those are secondary. The difference is about how much probability you assign to SpaceX successfully executing across Starship commercialization, Starlink Mobile, durable high-margin AI compute contracts, Terafab chip manufacturing, and orbital data centers, all within a 10-year window, while managing $29 billion in debt, a governance structure that limits minority shareholder protection, and competition from the most well-capitalized technology companies in history.

If you assign high probability to all of those outcomes simultaneously, $135 may be defensible. If you weight them as the probability-adjusted outcomes that the engineering timelines and capital requirements suggest, $44 is where the math lands.

14.Three Metrics to Watch After the Debut

1. Starlink ARPU and subscriber growth, reported quarterly. The connectivity business is the anchor of current intrinsic value. ARPU has already compressed from $99 in 2023 to $66 in Q1 2026. If ARPU stabilizes above $60 while subscribers continue growing at 20%+ annually, the Starlink bull case remains intact. If ARPU falls below $55 while subscriber growth decelerates, the base case for the entire company weakens materially.

2. AI segment gross margin and contract renewal economics. The Anthropic contract ($1.25 billion per month) and Google agreement ($920 million per month) are the two data points that most change the AI segment's near-term economics. The key question is not whether the revenue is real — it is — but whether the gross margin is 30%, 15%, or 5%. When SpaceX begins reporting quarterly segment financials as a public company, the AI gross margin will be the single most important number for long-term valuation.

3. Starship commercial launch cadence and cost-per-kg. Every element of the bull case — orbital compute deployment, Starlink V3 satellite launches, next-generation launch economics, and the reduction in Space segment R&D expense — depends on Starship achieving reliable, high-cadence commercial operations. The target is 100 metric tons reusable payload per launch. Track the number of commercial Starship launches per quarter and the reported cost-per-kg versus the Falcon 9 baseline of approximately $2,700 per kg.

15.A Final Observation

SpaceX is going public the night before what may be the most consequential decade in the history of space exploration, satellite communications, and AI infrastructure. The company has earned its place in that story through genuine engineering achievement — reusable rockets, a global satellite broadband network, and a vertically integrated physical stack that no competitor has yet replicated.

But the investor's question is never whether the company is extraordinary. The investor's question is always whether the price is right. At $135 per share, the market is asking you to pay today for Starlink's proven economics, launch dominance, Starship's potential, AI compute contracts, Terafab's ambitions, orbital data centers, and a governance premium for Elon Musk's vision — all at once, all at full probability. The DCF says that when you probability-weight those outcomes honestly and discount them at a rate that reflects the genuine risk of each, the answer is $44.23.

The gap between $44 and $135 is not a rounding error. It is the price of the story.

This analysis is for informational and educational purposes only. It is not investment advice, and nothing in this report constitutes a recommendation to buy, sell, or hold any security. The assumptions embedded in any DCF model reflect the analyst's judgment and can be wrong. You should do your own research, consult qualified financial advisors, and make your own independent investment decisions. The complete Excel valuation model, with all segment assumptions, cash flow projections, and sensitivity scenarios, accompanies this report and is available through the report delivery page and email.

DBOT is an automated AI research tool and does not provide investment advice. This report is generated for informational and educational purposes only and is not an offer, solicitation, or recommendation to buy or sell any security. AI-generated analysis may contain errors or omissions. You are solely responsible for your investment decisions — do your own research and consult a licensed financial advisor.