SPCX
A Generational Franchise, Priced for the Dream

Snapshot: Executive Summary

SpaceX is, on the merits, the best hardware franchise of the decade. Its May 20, 2026 Form S-1 disclosed audited financials for the first time: FY2025 revenue of $18.67B (+33%), split roughly Starlink/connectivity 61%, launch 22%, and the xAI/X AI unit 17%, with +$6.6B adjusted EBITDA¹. On June 12, 2026 it priced the largest IPO ever at $135 and closed the first day +19% at $160.95 (~$2.09T), with an intraday high near $176 (~$2.3T)². Since then the stock spiked to ~$226, then round-tripped below its $135 offer price; its first post-IPO print (Q2 2026, reported Aug 4) showed revenue of $7.81B (+92% YoY) and a much narrower $(541)M net loss, though $18.4B of quarterly capex spooked the tape⁸. Starlink is a genuine cash machine and launch is a near-monopoly. So why only a Hold?

Because you are paying for all of it, and then some:

• The multiple still prices the dream. ~$1.7T on $18.67B of FY25 revenue is ~92x trailing sales — down from ~112x at the first-day close, still roughly 2.5x Nvidia's price/sales — and ~2.2x the ~$780B independent fair value Morningstar published off the same S-1⁶. After the round-trip below the $135 offer, the tape now sits just under the offering's own rich ~$1.75T mark.
• Two businesses, one number. The S-1 is recast consolidated with xAI/X after the Feb 2026 merger, so the FY2025 $(4.9)B GAAP net loss blends a profitable space business with a ~$6.4B xAI operating loss. The Starlink segment alone earned ~$4.4B operating profit in FY25; the consolidated entity burned ~$14B of free cash flow¹. Q2 2026 narrowed the consolidated net loss to $(541)M as the AI unit's losses shrank relative to revenue, but $18.4B of quarterly capex kept it deeply FCF-negative.
• The unproven pillars gate the upside. Starship's reusability, Raptor reliability, and orbital refueling — the gates on both the ~$4.5B Artemis lunar contract and Starlink V3 economics — remain unproven; Flight 13 (Jul 24) deployed its first operational payload but the booster still hard-splashed⁵.

Our base-case fair value is $135 (Hold) — simply the IPO's own ~$1.75T mark. After the round-trip, spot (~$133) sits within ~1% of it, so the stock is now roughly at fair value rather than 19% above it. The risk/reward is more balanced but still skewed: Bull $177 (~$2.3T, ~+33%) needs Starship + Starlink V3 to scale; Bear $60 (~$780B, Morningstar) is a ~55% air-pocket if xAI burn, capex intensity, ARPU compression, competition, and political risk bite. We would move to a Buy on a credible path to xAI cash-burn narrowing plus Starship demonstrating booster catch and orbital refueling. Own this generational franchise — the price finally lets you, but the asymmetry still says wait.

Tactical: SPCX at ~$133 has round-tripped to just below our $135 base fair value (~2.2x Morningstar's ~$780B mark). This is a "great franchise, now at fair value" Hold, not a short: the Starlink engine and launch monopoly are real and compounding (Q2 Starlink +66%), and the Q2 loss narrowed sharply, but ~$18B of quarterly capex and the asymmetric bear keep us patient. The variables that decide the re-rate are xAI cash burn, capex intensity, Starship execution, and Starlink ARPU.

Investment Thesis

Probability weights 25% Bull / 45% Base / 30% Bear blend to a fair value of ~$123 — ~7% below the ~$133 recent close. The distribution is asymmetric: the bull is ~+33% while the bear is a ~−55% air-pocket to independent fair value. That skew, not any doubt about the franchise, is why we rate SPCX a Hold and would accumulate into the low-$100s and below.

Two SpaceXs Under One Ticker

The single most important thing to understand about SPCX is that the S-1 is recast consolidated with xAI/X under common-control accounting after the February 2026 merger. So the reported entity blends a profitable space/satellite business with a cash-torching AI unit. FY2025 revenue of $18.67B splits roughly Starlink/connectivity $11.4B (61%), space/launch ~$4.1B (22%), and the AI unit $3.2B (17%)¹. Revenue has compounded from ~$10.4B (2023) to ~$14.0B (2024) to $18.67B (2025). In Q2 2026 the mix shifted sharply as xAI scaled: Connectivity $4.29B (55%), AI $2.56B (33%), and space/launch $0.96B (12%)⁸.

The practical read: "SpaceX loses $4.9B" and "Starlink is a cash machine" are both true. Backing out the ~$6.4B xAI operating loss, the underlying SpaceX/Starlink business is roughly profitable (a derived ~$1.8B), and the Starlink segment alone earned ~$4.4B operating profit (~$7.2B EBITDA, ~63% margin). Valuing SPCX means deciding how much to pay for a world-class space franchise net of an AI unit that is, for now, a large negative. Q2 2026 showed the pattern holding but improving: the consolidated net loss narrowed to $(541)M as AI revenue (+247% YoY) outgrew its losses.

Starlink: The Profit Engine

Starlink is the reason SPCX is a mega-cap. Revenue reached $11.4B in FY2025 (+48%) and $4.29B in Q2 2026 (+66% YoY) at ~$1.66B segment operating profit, on 12.0M subscribers (doubled year-over-year) as of June 2026, up from 2.3M at the end of 2023³. It runs the largest satellite constellation ever assembled: ~10,880 spacecraft in orbit, more than half of all active satellites. The segment is FCF-positive at a ~63% EBITDA margin — a genuine broadband utility with a physical moat competitors are years from matching.

The catch is ARPU. Blended ARPU compressed from ~$99/mo (2023) to ~$66/mo (Q2 2026) as growth shifted to lower-priced tiers and cost-sensitive geographies — so revenue grows meaningfully slower than subscribers, and a +$10/mo hike in May 2026 only partly offsets it. The next legs are Direct-to-Cell (650+ DTC satellites; T-Mobile's "T-Satellite" live since Jul 2025, adding data; voice/broadband via V3 satellites slated for 2H2027) and Starshield, the government/defense arm (~$3B in 2025, including a ~$2.29B Space Force MILNET award). The bull needs the mix and defense revenue to out-run ARPU compression.

Launch: The Monopoly

The launch business is a genuine near-monopoly and the moat under everything else. SpaceX flew a record 165 orbital launches in 2025 — a sixth straight annual record, roughly one every 2.2 days, all reusable Falcon — capturing ~50% of global launches by count and >80% of all mass to orbit⁴. About three-quarters of those flights carried its own Starlink satellites, a vertically-integrated flywheel no competitor can replicate. Guidance is a deliberate slight slowdown to ~140-145 launches in 2026 as resources shift to support Starship operations. In Q2 2026 the space/launch segment posted $0.96B of revenue (+29% YoY) but a $(542)M operating loss as Starship development costs load into it.

Reuse economics are the edge: a ~$62M list-price Falcon 9 flies on an estimated ~$15M marginal cost (refurb ~$1M, propellant ~$0.2-0.3M; the booster is ~60% of vehicle cost). On top sits a deep, diversified government backlog: NSSL Phase 3 Lane 2 (~$5.9B, ~28 of 54 missions), NASA Commercial Crew (~$4.9B through ~2030), and Cargo Resupply (~$4.8B). Challengers are finally reaching orbit — Blue Origin's New Glenn is operational (first booster reflight April 2026) and cleared for top national-security missions, with Rocket Lab Neutron and China's reusable programs behind — but none is within an order of magnitude of SpaceX's cadence.

Starship: The Optionality

Starship is the call option embedded in the valuation. On July 24, 2026, Flight 13 — the second flight of the new V3 vehicle and the first since the IPO — deployed 20 operational Starlink V3 satellites (its first operational payload) and achieved the best reentry and splashdown to date, but the Super Heavy booster relit only ~10 of 13 landing-burn engines and hard-splashed, and a July 16 attempt aborted on Raptor turbo-pump failures⁵. The tower "catch" — the reusability endgame — has not yet been demonstrated on the new vehicle.

Why it matters to the model: V3 targets >100t to LEO and ~60 Tbps per launch (versus ~3 Tbps for a Falcon 9 Starlink flight), and it is the sole enabler of the Starlink V3 satellites the next revenue leg depends on. It is also the vehicle behind the ~$4.5B NASA Artemis Human Landing System contract — whose lunar landing has slipped from mid-2027 toward 2028+ and which requires undemonstrated orbital refueling (an estimated 10-16 tanker launches per Moon mission). Reusability, Raptor reliability, and orbital refueling are the three unproven pillars gating both Artemis and Starlink V3 economics. The bull underwrites them working; the bear underwrites the delays.

Financials (S-1)

The May 2026 S-1 is the first hard financial look at SpaceX. FY2025: revenue $18.67B (+33%), +$6.6B adjusted EBITDA, but a $(4.9)B GAAP net loss — the entire loss (and more) is the xAI unit's ~$6.4B operating loss. The Starlink segment earned ~$4.4B operating profit. This is an extremely capital-intensive enterprise: FY2025 capex was $20.7B (~$12.7B of it AI), free cash flow was roughly $(14B), cash fell $24.7B to $15.9B in Q1 2026, and long-term debt is ~$29.1B¹. The first post-IPO quarter (Q2 2026, reported Aug 4) showed the model inflecting: revenue $7.81B (+92% YoY), adjusted EBITDA $3.5B (+191%), and a much narrower $(541)M GAAP net loss ($(0.09) diluted) as the AI segment's losses shrank relative to revenue. But capital intensity is undiminished — Q2 capex was $18.4B (~$15.8B AI infrastructure); the post-IPO raise lifted cash to ~$100B against $47.5B of backlog⁸.

A critical caveat on every figure: the S-1 is recast consolidated with xAI/X, so standalone SpaceX profitability (~$1.8B in 2025) is derived, not cleanly disclosed. Two common conflations to avoid: (1) the Starlink segment profit (~$4.4B) is not company-wide net income; (2) leaked pre-S-1 "standalone SpaceX" numbers do not tie to the recast consolidated statements. The final column is a single quarter (Q2 2026), not a full year — SpaceX's first post-IPO quarterly, with segment splits now disclosed.

Valuation

There is no meaningful P/E — the consolidated entity loses money — so SPCX trades on sales and asset value. At the ~$133 recent close (~$1.7T) it changes hands at ~92x trailing (FY25) sales — down from ~112x at the first-day close — roughly 2.5x Nvidia's price/sales and still a large premium to every mega-cap. On the annualizing Q2 run-rate (~$31B) the forward multiple is lower (~55x) but still rich. Morningstar's independent fair value off the same S-1 is ~$780B — roughly half the IPO mark and ~2.2x below the current level⁶. With only a ~4% public float (the first lock-up expired Aug 7), price discovery is thin and volatile.

The valuation question is not whether SpaceX is great — it is whether ~92x sales for a capital-intensive, GAAP-lossmaking business (however extraordinary) leaves any margin of safety. Our answer is that spot has now round-tripped to the offering's own ~$1.75T mark — fairly valued on the base case, with the upside earned only if Starship and Starlink V3 deliver.

EV/Sales Scenario Model

Because SpaceX is loss-making on a consolidated basis, there is no meaningful P/E — so we anchor on revenue × an EV/Sales multiple, subtract net debt (~$13.2B), and divide by ~12.96B shares. We use trailing FY2025 sales ($18.67B) and calibrate the three multiples so each scenario reconciles to an externally-referenceable equity mark: the IPO's implied ~$1.75T (base), the first-day intraday peak ~$2.3T (bull), and Morningstar's independent ~$780B fair value (bear). The model retains the S-1 balance sheet (net debt ~$13.2B at Q1 2026) for continuity with the IPO-mark calibration; the post-IPO raise since lifted Q2 cash to ~$100B — a net-cash position — which we treat as upside to the framework (see Financials). Toggle Base / Bull / Bear / Reset and drag the sliders; the per-share value recomputes live.

Sensitivity: fair value / share ($) vs FY2025 revenue × EV/Sales

Fair-value calculator (quick cross-check)

Risk / Reward calculator

Note: The assistant reasons from the dashboard's data snapshot and thesis sections. It does not browse the web or access real-time fundamentals beyond what is in data.js. Treat its responses as scenario-modeling support, not primary research. Author judgments on rating, PT, and probabilities remain with the analyst.

The blended fair value (~$123) sits below spot, so the honest 12-month base view is $135 (the IPO's own mark) with a downward skew. Buy-case triggers: xAI cash-burn narrowing, Starship demonstrating catch + orbital refueling, and Starlink revenue re-accelerating despite ARPU compression.

Risks

• Key-person & political concentration (high). A June 2025 Trump-Musk feud triggered threats to cancel contracts and a White House-ordered review of SpaceX's ~$22B federal book. The valuation rests heavily on one politically volatile founder and a single dominant government customer⁷.
• Consolidated GAAP losses / xAI cash burn (high). The merger blends a profitable space business with a cash-torching AI unit: FY25 net loss $(4.9)B, Q1 2026 $(4.3)B, ~$(14B) FCF on $20.7B capex. Q2 2026 narrowed the net loss to $(541)M but capex was $18.4B in the quarter alone; the post-IPO raise lifted cash to ~$100B against ~$29.1B of long-term debt.
• Rich valuation vs fundamentals (high). ~92x trailing sales, ~2.2x Morningstar's ~$780B fair value; only ~4% float amplifies the volatility both ways (the stock has already round-tripped from ~$226 to below its $135 offer).
• LEO competition & ARPU compression (medium). Amazon Leo (best-capitalized rival, 2026 commercial launch) plus China's Guowang/Qianfan, against Starlink ARPU already compressed from $99 to $66/mo.
• Starship technical execution (medium). Booster catch, reusability, Raptor reliability, and orbital refueling are unproven and gate both Artemis (slipped to 2028+) and Starlink V3 economics.
• Launch challengers reaching orbit (medium). Blue Origin New Glenn is operational and cleared for national-security missions; Rocket Lab Neutron, ULA Vulcan, and Chinese reusables follow. None near SpaceX cadence yet, but the moat is narrowing.
• Regulatory & orbital-debris scrutiny (low). The FCC capped Gen2 at 7,500 satellites with a 5-year deorbit mandate; a March 2026 in-orbit fragmentation event raised debris and spectrum scrutiny.