Private Wealth · Investment Committee|24 July 2026|Strictly Confidential
Should we buy
SpaceX?
Space Exploration Technologies Corp. NASDAQ: SPCX — six weeks public, 49% off its peak,
and the most expensive large-cap in market history. An underwriting, not a vibe check.
Executive Summary
The answer: not at this price. Build the position on weakness, on our terms.
The business is real
- 80%+ of global mass to orbit; 165 Falcon launches in 2025, 99%+ success across ~650 flights.
- Starlink: $11.4B revenue, 63% segment EBITDA margin, 10.3M subs (Mar 31) — the fastest infrastructure monetization ever built.
- $27.8B/yr contracted AI compute (Anthropic, Google) landed within months of the xAI merger.
The price is not
- ~80x trailing revenue, ~190x trailing adj. EBITDA — ~4x Nvidia's peak AI multiple, at a $1.5T base.
- FCF −$13.8B (2025); S&P sees negative FCF through 2029. GAAP loss $4.9B.
- The 18-mo march $350B→$1.77T was multiple expansion — revenue grew 43%, the mark grew 5x.
- Lockup begins mid-August; ~44% of shares reportedly eligible by September.
The plan
Do not buy at $114. Our probability-weighted 2030 value is $101/share — you are paying today for the average of outcomes and wearing all the variance.
Ladder in through the lockup window: 0.50% at $80 · 0.75% at $70 · 0.75% at $60. Max 2.0% of portfolio at cost.
Pre-committed triggers to buy sooner, and to exit, on slide 17.
Prices/caps: CNBC SPCX quote & Bloomberg, Jul 15–24 2026 · Financials: SPCX S-1 (SEC, filed May 20 2026) via Payload, Via Satellite · FCF/ratings: S&P via Forbes (Cohan), Jul 10 2026 · Lockup: Axios Jul 17 2026; 44% figure single-source (Forbes) · Weighted value: our model, slides 12–16.
The Business|01
You are buying three companies stapled together
under one founder with 82% of the votes.
Connectivity — Starlink + Starshield
$11.39B +50% y/y
2025 revenue · +$4.4B operating income (39% margin) · ~$7.2B segment EBITDA (63%)
A satellite ISP with cable-company margins and no terrestrial buildout. 61% of company revenue. This is the profit engine — and the only segment that makes money.
Space — Launch + Dragon
$4.09B −$657M op
2025 revenue · loss absorbs ~$3B/yr Starship R&D ($930M in Q1'26 alone)
A near-monopoly deliberately run at a loss: launch exists to feed Starlink and fund the Starship bet. NSSL Phase 3 ($5.9B), Commercial Crew ($4.9B), HLS ($4.0B) anchor the book.
AI — xAI / Grok / compute
$3.2B −$6.4B op
2025 pro-forma revenue · $12.7B capex (60% of company total) · merged Feb 2 2026
An AI lab and compute-rental business injected into the company four months before the IPO. Carries $26.5T of the S-1's claimed $28.5T TAM — the load-bearing number in the valuation.
ConnectivitySpaceAIFY2025 segment operating income, $B — hover for detail
All figures: SPCX S-1 (SEC EDGAR, May 20 2026), segment disclosures, via Payload Research & Via Satellite S-1 coverage · Contract values: NASA, Space Systems Command releases · Voting: S-1/A Jun 3 2026 (~82% Musk voting power via 10:1 Class B, 40.5–42% economics).
The Business|02
Eighteen months, a 5x mark-up, and a 49% drawdown.
The repricing was multiples, not fundamentals.
Company valuation mark, $Btenders → merger → IPO → public trading · hover any point
Valuation Dec '24 → IPO
5.1x
$350B → $1.77T in 18 months
Revenue over same period
+43%
FY24 $14.0B → FY25 $18.7B (S-1)
Implied EV / trailing revenue
27x → 95x
Dec '24 tender → IPO pricing. Today: ~80x. Every turn of that expansion is borrowed from the future you're underwriting.
Marks: Bloomberg Dec 10 2024 ($350B, $185/sh) · tender recaps Jul 2025 ($400B) · SatNews Dec 6 2025 (~$800B) · CNBC Feb 3 2026 (xAI merger, $1.25T combined) · TechCrunch Jun 11 2026 (IPO $135, $1.77T) · Forbes/CNBC Jun–Jul 2026 (peak $225.64 Jun 16 ≈ $2.96T; $114 Jul 24 ≈ $1.50T). Multiples: our calculation on S-1 revenue — EST
Business Unit · Launch|03
Launch is a monopoly run at breakeven —
and three quarters of it is SpaceX launching for itself.
Starlink (internal, no revenue)Customer missionsFalcon launches per year
Share of global upmass '25
>80%
Share of launches '25
~50%
Success, ~650 flights
99%+
Record booster reuse
36 flights
Unit economics per Falcon 9
List price $74M (2026) vs ~$15–28M internal marginal cost EST — Musk (Aviation Week, 2020) $15M best case; Couluris $28M all-in. A ~70% gross-margin product sold mostly to itself.
Why it still loses money
The Space segment posted −$657M operating income in 2025 because it carries ~$3B/yr of Starship R&D — >$15B cumulative. Launch is priced as a moat and an internal subsidy, not a profit center. The contracted federal book (NSSL-3 $5.9B · Crew $4.9B · HLS $4.0B · Golden Dome $4.2B) is the floor under it.
Launch counts: SpaceNews/NASASpaceflight year-end roundups; Space.com Jan 2026 (165 Falcon, 123 Starlink) · Upmass/share: BryceTech via Via Satellite Apr 10 2026 · Pricing: SatBase Feb 2026 ($74M, $7,000/kg rideshare) · Costs: ElonX (Musk, Aviation Week May 2020), press-call $28M · Starship spend: Bloomberg May 20 2026 · Contracts: SSC Apr 2025, NASA, Bloomberg May 29 2026.
Business Unit · Starlink|04
Starlink is compounding subscribers at 2x/year —
while cutting its own prices 18% a year to do it.
Subscribers, millionscompany disclosures · Jun '26 point is tracker data EST
Blended ARPU, $/month — S-1
S-1 states ARPU will keep declining: international mix, cheaper tiers, deliberate capacity-filling.
Segment EBITDA margin
41% → 50% → 63% ('23→'25). Vertical launch integration is the structural cost advantage no competitor has.
Subscribers: company milestones via Erlichman compilation; S-1 (2.3M YE23 · 4.4M YE24 · ~9.2M YE25 · 10.3M Mar 31 '26); 12M Jun '26 per trackers (Yahoo Finance) · ARPU: S-1 ($99→$91→$81→$66 Q1'26), Roic Apr 29 2026 · Margins: S-1 segment data via New Space Economy May 30 2026 · YE26 forecast: Quilty Space (16.8M subs).
Cost Structure|05
The capital treadmill: capex quadrupled in a year,
and the AI lane now outspends the rockets 3-to-1.
AI / data centersSpace (Starship, pads, satellites)Capital expenditure, $B
Q1 2026 bar shows the quarter annualized (×4): $10.1B in one quarter, $7.7B of it AI. 2024 split not disclosed — shown as Space EST.
Free cash flow, FY2025
−$13.8B
S&P projects negative FCF through 2029. Debt: $29.1B principal + $25B post-IPO bond already trading at junk-like spreads (single source — Forbes).
The satellite treadmill
~$2–3B/yr EST
~5-year satellite life; 1–2 deorbits/day already (472 in six months). At 30k satellites this is permanent maintenance capex wearing a growth-capex costume.
War chest vs burn
~$91B vs ~$40B/yr
IPO proceeds ($75B) + Q1 cash ($15.9B) against the Q1'26 capex run-rate. The IPO wasn't a liquidity event — it was a financing event. That tells you what management thinks the spend curve looks like.
Capex: prospectus ($5.6B '24 · $20.7B '25 · $10.1B Q1'26, $7.7B AI) via TNW/Bloomberg, HL, Yahoo · FCF: Forbes (Cohan) Jul 10 2026, leverageshares · Deorbits: McDowell via EarthSky/The Register Oct 2025; company disclosure Dec '24–May '25 · Treadmill math: ~2,000 sats/yr × ~$1.1M all-in — our estimate on third-party unit costs.
Business Unit · AI|06
The AI segment: $27.8B of contracted revenue —
from three customers, on 90-day termination clauses.
Annualized contracted compute revenue, $B — hover for terms
Largest customer share
54%
Anthropic: $1.25B/mo through May 2029
Termination notice
90 days
either party, both anchor contracts
2025 AI op loss
−$6.4B
on $3.2B revenue; $12.7B capex
Read this the way a lender would
- The growth is real: $3.2B (2025) → ~$27.8B contracted run-rate in under a year. Colossus is 1GW+ and filled.
- The quality is not: two customers are 94% of it, both can walk in 90 days, and both are building their own capacity. Anthropic rents here while competing with Grok — this is overflow demand, not committed demand.
- The S-1's $26.5T AI TAM is the number doing the valuation work. Damodaran rejects the framing outright; Morningstar carries the AI business at ~$170B probability-weighted.
- Orbital data centers ("AI1", ~2028) are the Starship-dependent sequel — a real option, not a plan you can underwrite.
Contracts: TechCrunch May 20 2026 (Anthropic $1.25B/mo thru May 2029, 90-day outs), TechCrunch Jun 5 2026 (Google $920M/mo), Yahoo Finance Jul 2026 ($27.8B total, 3 customers) · Segment P&L: S-1 via Payload/HL · TAM: S-1; Damodaran (Apr 23 2026); Morningstar Jun 2026 · Orbital DCs: S-1/Fortune Jul 19 2026.
Execution|07
Starship is the load-bearing wall: Starlink V3, orbital compute, and the whole bull case sit on it.
✓ Success
◐ Partial
✕ Vehicle lost
Integrated flight tests 1–13 · hover for outcome
2025 cadence
5 of 25
flights flown vs FAA-authorized
Sunk development
>$15B
disclosed in prospectus
Cost per V3 stack
~$90M EST
target: full reuse, $10–100/kg
Musk-time multiplier
~2x
historical slippage on Starship milestones
What hangs on it
- Starlink V3 satellites (~2,000kg, 6x the capacity) physically don't fit on Falcon 9. No Starship cadence → no capacity growth → the congestion bear case wins.
- Orbital data centers — the entire space-AI story — need 100t reusable lift at airline cadence.
- HLS / Artemis: $4.0B contracted; NASA OIG already flags schedule risk.
- Flight 13 is now — first 20 real V3 Starlink sats. V3 flew once (Flight 12, May '26): ship survived reentry, booster was lost. Progress is real; "airline cadence" is not yet.
Flight record: Wikipedia/NASASpaceflight/Spaceflight Now flight pages (F1–F12), Space.com Jul 2026 (F13) · Cadence authorization: FAA Tiered EA May 2025 (25/yr Boca Chica), Feb 2026 (44/yr Florida) · $15B/$90M: Bloomberg May 20 2026, TNW prospectus coverage · Slippage: Musk statements 2016–2025 vs actuals (Al Jazeera May 2025, OpenTools).
Competitive Position|08
The moat is real and measured in years —
but every element of it is now being replicated.
LEO broadband constellations — satellites in orbit, July 2026 · hover for trajectory
Scale disparity is the story: Starlink has 28x its nearest live competitor and launches its own replacements. But Kuiper's binding FCC deadline was waived (Jul 2026) — Amazon can now lose money on Leo indefinitely.
Erosion vectors, honestly weighed
- Blue Origin New Glenn: 3 flights, 2 sea landings, first booster reflight Apr 2026. Western reusability is no longer unique — cadence still ~2–3/yr.
- Amazon Leo: ~390 sats, enterprise beta live, consumer service late 2026; AWS channel + balance sheet. 18–24 months behind Starlink's curve at scale.
- China (Qianfan + Guowang): ~400 sats combined, 28k planned; Zhuque-3 reached orbit Dec '25. Locks SpaceX out of ~1/3 of the addressable planet — a TAM haircut, not a competitor.
- AST SpaceMobile: the D2C broadband spec-leader; service activation H2 2026. Watch it — Starlink paid ~$19.6B for EchoStar spectrum to hold this flank.
- ULA / Ariane / Neutron: re-securing sovereign payloads; no price threat before 2028.
Constellations: HighSpeedInternet Jul 23 2026 (Starlink ~10,850 in orbit); CNBC Jul 2 2026 (Leo ~390); china-in-space.com Jul 4 2026 (Qianfan 218, Guowang ~190 EST); Space Intel Report (OneWeb ~630) · New Glenn: Blue Origin releases; Space.com/Fortune Apr 19 2026 · Kuiper waiver: FCC via IndexBox 2026 · EchoStar: 8-K Sep 8 2025 (~$17B) + SpaceNews Nov 2025 (+$2.6B) · AST: satelliteinternet.com 2026.
The Case Against|09
The steelman bear: two unproven miracles, multiplied together, at 80x sales.
- Physics caps the core business. Starlink's capacity per cell bounds it to rural/mobility niches — Morningstar sizes the real TAM at $129B, not $1.6T. Evidence exists: US median speeds fell 91→63 Mbps in 2022 as cells filled; Indonesia fell ~2/3 within a year of launch. Escaping the ceiling requires Starship-scale upmass — miracle #1.
- The AI revenue is concentrated and callable. 94% from two customers on 90-day outs, both building competing capacity in-house. Monetizing the $26.5T TAM claim requires orbital data centers — miracle #2, which also requires miracle #1.
- The cash math is hostile. −$13.8B FCF, negative through 2029 (S&P); $54B of debt+bonds; a satellite fleet that must be rebuilt every 5 years forever.
- Governance is pre-broken. 82% supervoting, controlled-company exemptions, a pre-IPO related-party merger (xAI) that moved $250B of Musk's private exposure onto public shareholders, and a CEO who threatened to decommission Dragon during a feud with the President. CII and NYC/NYS comptrollers objected in writing.
- Supply is coming. Lockups open mid-August; ~44% of shares eligible by September (single source). The stock already broke IPO price with insiders still locked.
Where serious independent work lands
| Valuer | Implied/share | vs $114 |
|---|
| Morningstar fair value | $63 | −45% |
| Damodaran intrinsic (median) | ~$95 | −17% |
| Goldman Sachs target | $205 | +80% |
| Morgan Stanley base | $300 | +163% |
| Raymond James | $800 | +602% |
A sell-side range of $75–800 is not price discovery — it is an admission that nobody can underwrite this yet. The two valuers with no banking relationship sit below the price.
The historical rhyme
Iridium spent $5B, had a working global constellation, and filed Chapter 11 nine months after launch — the assets sold for 0.7¢ on the dollar. The infrastructure was real; the first-wave equity still died. The counter: Starlink has 10M+ paying subscribers at 63% margins. Iridium died at 20,000. Real revenue is the difference — price is the question.
Morningstar: "Why We Think the SpaceX IPO Is Overvalued," CNBC Jun 3 2026 · Damodaran: substack/blog Apr 23 2026 ($1.22–1.29T ÷ 13.1B sh — per-share is our conversion EST) · Targets: Bloomberg Jul 7 2026 initiation wave · Speeds: Ookla via Fierce/katadata · Governance letters: NYC Comptroller May 13, CII Jun 9 2026 · Iridium: Tuck case study, Smithsonian.
Our Model|10
The forecast, with the assumptions on the table.
Three self-consistent 2030 worlds. Every driver below is anchored to a disclosed 2025 actual; nothing is free-floating. Toggle the scenario — the table and the next slide move together. ALL FORWARD FIGURES ARE OUR ESTIMATES
2025 anchors: S-1 segment disclosures. Scenario logic — Bear: congestion + Kuiper price war + Anthropic exercises its 90-day out in 2027 + Starship slips again. Base: Starlink lands mid-consensus (35M subs vs street 50–80M), AI contracts hold and broaden modestly, Starship reaches useful-but-not-airline cadence. Bull: Starship ≥50 flights/yr unlocks V3 capacity + D2C/EchoStar spectrum monetizes + AI book triples off contracted base (still half of Morgan Stanley's 2030 AI number).
Our Model|11
Revenue builds to $95B by 2030 in the base case.
ConnectivitySpaceAIRevenue by segment, $B · 2025 actual, 2026–30 modeled · hover
2025 actuals: S-1. All 2026–2030 bars are our model output under the selected scenario (drivers on previous slide). For calibration: Quilty 2026E ~$20B total revenue / $14B Starlink EBITDA; Morgan Stanley 2030E Starlink $144B (we are far below it in every scenario); ARK 2030E ~$200B total.
Valuation|12
What $114 already assumes: our bull case, delivered on schedule.
EV / revenue across the comp set — SPCX highlighted · trailing unless noted · hover
The entire space complex re-rated to 20–130x sales in 2026. The last time small-cap space traded like this (2021 SPACs), ~$50B of paper value was destroyed and the median name fell 85–97%.
The reverse math
For SPCX to return 10%/yr from $114, it must be worth ~$2.4T by end-2030. At a generous 30x exit EV/EBITDA, that requires ~$80B of 2030 EBITDA — 2.2x our base case ($37B), 12x what the company earned in 2025, and within rounding distance of our full bull case.
Paying a price that needs the bull case is how you turn a great company into a bad investment.
Multiple context
| EV/Rev | EV/EBITDA |
|---|
| SPCX today | ~80x | ~190x |
| Nvidia at its AI peak | — | ~70x |
| Mature telco/cable | ~1–2x | ~5–7x |
Somewhere between those rows is where Starlink's terminal multiple lives. The debate is which row it's closer to.
Comps: Motley Fool Jul 16 2026 (ASTS >130x fwd, RKLB ~60x fwd), stockanalysis/GuruFocus Jul 2026 (PLTR ~60x), Planet ~21x fwd, Comcast ~1.1x (macrotrends) — mixed trailing/forward as labeled EST · Nvidia 70x: Seeking Alpha Jun 2026 · SPCX: our calc on S-1 financials at $1.50T · SPAC losses: Wolf Street, SpaceNews, CNBC 2024.
Our Model|13
Three worlds, one price. The market is charging you $114 for a $101 weighted ticket.
2030 per-share outcomes (our model) and today's external marks · hover any bar
Implied returns from $114
| Scenario · weight | 2030 | IRR |
|---|
| Bear · 25% | $10 | −42%/yr |
| Base · 50% | $74 | −9%/yr |
| Bull · 25% | $248 | +19%/yr |
| Weighted | $101 | −3%/yr |
The honest counter to ourselves
If our bull weights are wrong the way AWS bears were wrong in 2015 — if orbital monopoly + AI demand compounds past every ceiling — this table undercounts the right tail badly, and no entry discipline will matter. That is a real possibility. It is 25% of our book, not 100% — because the bull case needs Starship cadence and AI diversification and a capacity-ceiling escape to all land, and SpaceX's own 2025 cadence promise came in 80% short.
Model outcomes: our estimates (assumptions, slide 12; equity value = EBITDA × exit multiple ± net cash, ÷ ~14.0B fully-diluted 2030 shares incl. EchoStar/Cursor stock consideration & comp dilution) · External marks: Morningstar $63, Damodaran ~$95 (our per-share conversion), GS $205, MS $300 (bull $600/bear $75), Raymond James $800 — Bloomberg Jul 7 2026.
Our Model|14
Sensitivity: what you need to believe for $114 to be cheap.
Implied 2030 share price = (EBITDA × multiple) ÷ 14.0B shares. Blue cells beat today's $114; red cells lose money. The white ring marks our base case. Hover any cell for the IRR.
How to read it
Today's price only wins in the upper-right quadrant — EBITDA ≥ $65B and an exit multiple ≥ 30x. Our base case ($37B, 28x) lands at $74. Even $50B of EBITDA at 30x — a 7.5x improvement on 2025 — only gets you to $107. The margin of safety at $114 is negative across most of the plausible grid.
Where it flips
At our ladder prices ($80 / $70 / $60), the break-even EBITDA-multiple frontier drops by roughly a third — the base case turns from a −9%/yr hold into a +4 to +10%/yr hold, and the bull case pays 25–30%/yr. Same company, same forecast; the entry price is the entire difference.
Grid: our model. Share count held at 14.0B fully-diluted 2030E; net cash held at $0 (base) — bear/bull cash adjustments shown on slide 15 change terminal values by <±$4/share. All cells are estimates.
Discipline|15
Pre-committed: what makes us buy more, and what makes us walk.
Upgrade triggers — add ahead of the ladder, pay up to ~$100
- AI book diversifies: ≥6 compute customers with the largest <30% of contracted revenue, or any anchor contract converted to take-or-pay ≥3 years.
- Starship reaches cadence: ≥20 flights over any trailing 12 months with booster + ship reuse demonstrated, and V3 Starlink deployment running.
- Starlink holds price: YE26 subs ≥16M (Quilty's number) with blended ARPU stabilizing ≥$60 — growth without buying it.
- Governance concession: any sunset on supervoting, independent board majority, or a Starlink segment listing that gives public holders direct claim on the profit engine.
Exit / stand-down triggers — sell the position, whatever the loss
- Anthropic or Google exercises the 90-day out (or renegotiates >25% down): the AI segment reprices to its churn case and our bear scenario becomes base.
- Starlink net adds fall below ~1M/quarter for two consecutive quarters with ARPU still declining — the ceiling argument winning in the data.
- A Starship failure that grounds the program >12 months, or Artemis/HLS termination for convenience.
- Related-party escalation: another Musk-entity acquisition (Tesla, X follow-ons, Anysphere close >$60B) on terms public holders can't block.
- Political rupture with teeth: actual contract cancellation, not tweets — the June 2025 feud template, executed.
Calendar checkpoints
Aug–Sep '26: lockup expiries — expect supply, do not pre-empt the ladder · ~Sep '26: first full-quarter earnings (Q2) — first look at AI segment margin · Q4 '26: Kuiper consumer launch + YE Starlink subs vs 16.8M · May '29: Anthropic contract term.
Trigger thresholds: our judgment, anchored to Quilty YE26 forecast, contracted terms (TechCrunch May/Jun 2026), FAA authorizations, and the June 2025 Musk–Trump episode (SpaceNews/PBS) as the political stress template.
Recommendation|16
Own the company. Refuse the price.
Action today
Do not initiate at $114. No chasing, no FOMO exception.
Standing orders
Ladder to a 2% position through the lockup window — and let the supply come to us.
This is a venture bet wearing a mega-cap costume: a −42%/yr bear case and a +19%/yr bull case cannot be sized like a blue chip. At 2% of the portfolio, the bear case costs ~1.7 points of NAV; the bull case adds ~2.4. That asymmetry is acceptable only at our prices — which the August–September lockups are structurally likely to offer.
The ladder
| Limit | Size | Exp. IRR at entry* |
|---|
| $80 | 0.50% | +5%/yr |
| $70 | 0.75% | +9%/yr |
| $60 | 0.75% | +13%/yr |
| Blended (~$69) | 2.00% | ~+9%/yr |
*Probability-weighted, our model. GTC orders, working from Aug 1. Cap: 2.0% at cost, 4% at market — trim above.
House rules
- Horizon 5+ years; expect ≥50% drawdowns en route — pre-agree that now.
- Triggers on slide 17 override the ladder in both directions.
- If nothing fills by year-end and no triggers fire, we reassess at Q4 earnings — doing nothing is an acceptable outcome.
- Review at every checkpoint; this memo expires at the Q2 print.
Recommendation is the committee's own analysis for this client's growth sleeve; assumes diversified portfolio, no leverage, no options overlay. Not investment advice to any other party. Sizing math: 25% × (−85%) ≈ −1.7pts at 2%; 25% × (+118%) ≈ +2.4pts (cumulative 2030, our model).
Appendix
Sources & provenance
Company filings & primary
SPCX S-1 / S-1A (SEC EDGAR, May 20 & Jun 3 2026) — segment revenue & operating income, ARPU, subscribers, capex, debt, voting structure, TAM claims · SpaceX IPO pricing announcement (Jun 11 2026) · EchoStar 8-K (Sep 8 2025) — spectrum sale terms · FAA Tiered EA (May 2025), FAA Feb 2026 (Florida 44/yr) · NASA & Space Systems Command award releases (CCtCap $4.93B; NSSL-3 Lane 2 $5.9B; HLS $2.89B + $1.15B; Golden Dome $4.16B, May 29 2026).
Financial press
Bloomberg (Dec 10 2024 tender; May 20 2026 Starship $15B; Jul 7 2026 initiations; Jul 15 2026 IPO-price break) · CNBC (Dec 11 2024; May 6 & Jun 3/12 2026; Jul 2 2026 Leo) · TechCrunch (Jun 11 2026 IPO; May 20 2026 Anthropic $1.25B/mo; Jun 5 2026 Google $920M/mo) · Fortune (Dec 13 2025; Jul 19 2026 compute business) · Forbes/Cohan (Jul 10 2026 — debt, FCF, lockup; several figures single-source, flagged) · Axios (Jul 17 2026 lockup) · SatNews (Dec 6 2025, $800B tender).
Industry & independent research
Payload Research (S-1 teardown; launch stats) · Quilty Space (Starlink model; 2026E $20B rev / $14B EBITDA / 16.8M subs) · BryceTech via Via Satellite (Apr 10 2026 — launch & upmass share) · Morningstar ("Why We Think the SpaceX IPO Is Overvalued," Jun 2026, FV $63) · Damodaran (blog/substack, Apr 23 2026, $1.22–1.29T intrinsic) · ARK Invest (2030 EV model) · Morgan Stanley roadshow model via Roic (Jun 2026) · Ookla via Fierce Network/Broadband Breakfast/katadata (speed data) · McDowell/EarthSky (deorbit rates) · Tuck School Iridium case study.
Space trade press
SpaceNews · NASASpaceflight · Spaceflight Now · Space.com · Via Satellite · Space Intel Report · china-in-space.com (Chinese constellations) · ElonX (Musk cost statements, Aviation Week 2020) · Teslarati · European Spaceflight.
Estimate policy: every figure not in a company filing is tagged EST with its estimator named in the slide footnote. Known conflicts we did not resolve: Q1'26 net loss (−$1.94B vs −$4.28B across S-1 press coverage); Musk voting share (82.4% S-1/A vs 85.1% earlier coverage — we use the later filing); Jun '26 subscriber count (12M is tracker data, not company). Forward model figures (2026–2030) are entirely our own and appear only on slides marked "Our Model."
← → navigate · 1–9 jump · B/N/U scenario