Phase A — Understand the business
Lens 1 · Company Overview
SpaceX (Space Exploration Technologies Corp.) is now three businesses stapled together under one founder-controlled umbrella:
- Starlink (connectivity) — the cash engine. A low-Earth-orbit (LEO) broadband constellation selling consumer, enterprise, maritime, aviation, government, and (new) Direct-to-Cell connectivity. $11.4B revenue in FY25, +48% YoY, 61% of total, with $4.4B segment operating profit — the only profitable division. 10M+ active customers across 160 countries by Feb 2026; adding 750k–1.5M/month.
- Space / Launch Services — the moat factory. Falcon 9 / Falcon Heavy / Dragon for commercial + government payloads, plus the in-development fully-reusable Starship. $4.1B revenue FY25, +8% YoY, but a −$619M operating loss (Starship R&D dwarfs external launch revenue). 165 orbital launches in 2025, 6th consecutive annual record.
- AI (xAI / Grok / X) — the moonshot bolt-on. Absorbed via the Feb 2026 SpaceX↔xAI merger ($1.25T combined: SpaceX $1T + xAI $250B). $3.2B FY25 revenue but a −$2.5B operating loss. xAI was reportedly burning ~$1B/month.
Business model in plain terms: SpaceX uses the world's only high-cadence reusable launch system to deploy its own satellites at near-zero marginal launch cost, then sells the bandwidth on a recurring subscription. The launch business is barely break-even on its own; its real value is as an internal utility that makes Starlink possible at a cost no rival can match. The AI arm is a capital sink funded by the connectivity cash flows.
Key payment terms / structure: Starlink = recurring monthly subscription (ARPU fell 18% to ~$81/mo between 2023 and 2025 as the mix moved toward lower-priced international consumers). Launch = milestone/firm-fixed-price government contracts (NASA, DoD/Space Force) + commercial manifest. Customer concentration is meaningful on the launch side (Pentagon + NASA drive the $4.1B).
confirms the headline: FY25 revenue **$18,700M**, operating income **−$2,600M**, net income **−$4,900M**; the file is a deliberate "fill-from-S-1 by hand" scaffold (segments/balance sheet still blank). Cross-confirmed.
Lens 2 · Supply Chain
SpaceX is the textbook vertically-integrated manufacturer — ~85% of components produced in-house. This is the defining structural fact of the business and the reason the moat is so deep. The chain, with named stakeholders:
Upstream inputs → SpaceX:
- Additive manufacturing / metal 3D printing: Velo3D (VELO) — SpaceX bought ≥22 Sapphire laser powder-bed-fusion machines and signed an $8M licensing/support deal (Sept 2024) used heavily in Raptor engine development; SuperDraco thrusters are fully 3D-printed.
- Raw materials & specialty alloys: stainless steel (Starship airframe), aluminum-lithium, Inconel, carbon composites — sourced via 3,000+ external suppliers for parts not made internally.
- Avionics/electronics: SpaceX designs its own flight computers and avionics in-house; commodity silicon and passives bought externally.
SpaceX (manufacture):
- Engines: Merlin (Falcon), Raptor (Starship) — built and iterated in-house, the core of the cost advantage.
- Satellites: Starlink V2/V3 — historically 45+/week (~5–6/day) production; Direct-to-Cell sats carry an eNodeB modem on 4G LTE bands.
- User terminals: Bastrop, TX facility produced 1M terminals in its first 9 months; ~15,000 kits/day capacity cited.
SpaceX → end customer:
- Launch buyers: NASA, U.S. Space Force/DoD, commercial sat operators, and — critically — competitors (Amazon Kuiper/Leo, Guowang, Qianfan all depend on low-cost reusable launch, much of which only SpaceX can supply).
- Connectivity buyers: 10M+ Starlink subscribers; carrier partners for Direct-to-Cell.
Chokepoints / single-source dependencies:
- Inbound: Velo3D is a notable specialized single-source for some Raptor printing, but SpaceX's in-house posture limits external single-points-of-failure by design. The real chokepoint is internal — Raptor production rate gates Starship cadence.
- Outbound — SpaceX IS the chokepoint for the rest of the industry. It controls ~82% of the global commercial launch market; rivals deploying constellations are partly hostage to SpaceX's manifest and pricing. This is the single most important supply-chain fact: the company is a chokepoint, not a victim of one.
Lens 3 · Competitive Advantages (moats)
SpaceX has one of the widest moats in any industry — and it is structural, not brand:
- Cost moat (reusability × cadence): Falcon 9 lists ~$69.75M, roughly $2,700–3,000/kg (reused vs. expendable), against legacy expendable competitors at multiples of that. A booster has flown a record 29 times. No competitor has matched orbital-class reuse at cadence. This compounds: cheaper launch → cheaper Starlink deployment → more sats → more revenue → funds Starship → even cheaper launch.
- Scale / experience-curve moat: 165 launches in 2025 vs. low-single/double digits for everyone else. Manufacturing-rate leadership (engines, sats, terminals) is years ahead.
- Vertical integration (~85% in-house): controls iteration speed and cost; insulated from supplier hold-up.
- Switching costs / network effects (Starlink): the only at-scale LEO broadband with global coverage; first-mover spectrum + orbital slots; Direct-to-Cell adds a carrier-embedded layer with no SIM/app change.
- Regulatory/orbital moat: licensed slots and spectrum are scarce; SpaceX's incumbency (and its DC relationships) is a barrier. It has even asked regulators to deploy up to 1 million AI satellites for orbital data centers — an audacious land-grab on orbital real estate.
Bargaining power: Overwhelming over launch customers (it's the only high-cadence game in town, and it launches its competitors' constellations). Strong over suppliers (in-house posture + 3,000 vendors = low dependence). Moderate over Starlink consumers (ARPU is falling 18% as it chases lower-priced international subs — the one place its pricing power is visibly eroding).
The honest moat caveat: the moat is on launch + connectivity. The newly-bolted AI segment has an "economic moat indeterminate" per Morningstar — Grok vs. OpenAI/Anthropic is an open contest, and the $60B Cursor acquisition is an attempt to buy moat it doesn't organically have.
Lens 4 · Segments
segments.csv is empty (web-only). FY25 segment picture from the S-1 coverage:
| Segment | FY25 Revenue | YoY | Operating income | Share | Source |
|---|
| Starlink (connectivity) | $11.4B | +48% | +$4.4B | 61% | |
| Space / Launch services | $4.1B | +8% | −$619M | 22% | |
| AI (xAI/Grok/X) | $3.2B | n/a (acq.) | −$2.5B | 17% | |
| Total | $18.7B | +43% | −$2.6B | 100% | `` + |
Trend & cause:
- Starlink decelerating-but-dominant: +48% rev is the growth story, but the 18% ARPU decline (2023→2025) signals the high-margin US/Western-Europe residential market is mature; future growth is mix-shift to higher-ARPU enterprise/maritime/aviation/Direct-to-Cell. Q1 2026 connectivity already posted ~$1.19B segment profit.
- Launch flat (+8%): external launch revenue is structurally capped by how much of SpaceX's cadence is consumed internally by Starlink (63 dedicated Starlink launches in 2025). The −$619M op loss is Starship R&D eating the segment.
- AI is a brand-new, deeply loss-making line (−$2.5B) bolted on in Feb 2026 — it changes the company's risk profile materially and is the single biggest source of valuation disagreement.
Geography: not broken out in available sources — n/a at the segment×geo level. Starlink spans 160 countries; launch revenue is US-government-weighted.
Phase B — Measure performance
+private overlay note: Lens 5 → Funding & valuation trajectory; Lens 7 → Cap table & secondary marks; plus a Traction & unit-economics add. The recent IPO means these are partly "just happened" rather than forward-looking.
Lens 5 · Funding & Valuation Trajectory (overlay swap)
SpaceX's private-round march is one of the steepest in venture history, capped by the largest IPO ever:
| Date | Event | Valuation | Lead(s) | Source |
|---|
| Jan 2015 | Series F, $1B | $10B | Google, Fidelity | |
| Feb 2021 | $1.16B | $74B | Sequoia, Valor, Coatue, Fidelity | |
| Jun 2022 | $1.68B | $127B | Mirae Asset | |
| Jan 2023 | $750M | $137B | Andreessen Horowitz | |
| Dec 2024 | $750M round + $500M buyback | $350B | a16z | |
| Jul 2025 | tender/round | $400B | — | |
| Dec 2025 | tender/round | $800B | — | |
| Feb 2026 | xAI merger | $1.25T combined (SpaceX $1T + xAI $250B) | triangular merger | |
| Jun 12 2026 | IPO (SPCX, NASDAQ) | ~$1.77T at $135 offer; ~$2T+ on day-one close ($160.95, +19%) | raised ~$75B / 555M+ shares — largest IPO ever | |
`` anchors: IPO price $135, live ref $214.80 (2026-06-16) → ~$2.8T implied; annual implied vol ~129%. (Note: other sources put the June-18 print at ~$191.82 with the 12-mo analyst avg target $188.17, range $62–$310 — see Lens 8.) Burn signal: −$4.9B FY25 net loss against ~$11.9B total private capital raised across 30 rounds means the IPO's ~$75B raise is the new fuel tank.
Lens 6 · Founder/Operator Sentiment (overlay swap — no earnings calls yet)
No earnings calls exist (IPO'd 6 days ago). Tone from the IPO roadshow / management interviews:
- Gwynne Shotwell (President/COO) ran the investor-facing message — disciplined, operations-first, framing Starlink profitability and launch cadence as the proof points; signaled the Tesla relationship "makes Elon's life a little easier".
- Musk's framing is maximalist: a "vertically-integrated innovation engine on (and off) Earth" spanning AI, rockets, space internet, and X; publicly floated $1T revenue by 2030 (vs. Morgan Stanley's $330B) — the aspiration/fundamentals gap is the whole bear case in one number.
- Former Nasdaq chief Robert Greifeld: the stock trades "not on fundamentals" but "on the aspiration of what's possible with human spirit" — a striking thing for a market veteran to say about a $2T company.
What management emphasizes: Starlink as the proven cash engine; Starship as the next platform; AI/orbital-compute as the optionality. What's conspicuously underweighted: a credible near-term path to GAAP profitability and the dilution math of using stock for $60B acquisitions.
Lens 7 · Cap Table, Secondary Marks & Comps (overlay swap)
Cap table / control (the headline governance fact):
- Musk: ~42% equity, but ~82–85% of voting power via a dual-class structure (Class B = 10 votes; Musk holds 93.6% of Class B). He cannot be removed — it requires a Class B majority he alone controls.
- Board majority is not independent; no independent nominating/comp committee; a "Corporate Opportunities" clause lets Musk pursue deals via Tesla/Boring/Neuralink.
- Syndicate quality (IPO-proximity tells, now realized): Google, Fidelity, Sequoia, a16z, Coatue, Mirae, Valor — exactly the tier-1 + crossover roster that precedes a mega-IPO.
Comps (public space peers) — provenance-critical, multiples are `` or n/a:
| Company | Ticker | ~Market cap | EV/Sales | Notes | Source |
|---|
| SpaceX | SPCX | ~$2.0–2.8T | ~94x at IPO → ~110–150x live (on $18.7B FY25 rev) | only profitable segment is Starlink | `` + |
| Rocket Lab | RKLB | ~$49–72B | ~70x fwd rev (on ~$0.6B FY25 → ~$0.9–1.2B fwd) | reusable Neutron debut 2026 | |
| AST SpaceMobile | ASTS | n/a (mkt cap) | n/a | pre-revenue D2C; $1B run-rate target 2027 | |
| Planet Labs | PL | n/a (mkt cap) | n/a | $307.7M FY26 rev, +26% | |
| Blue Origin | private | private | n/a | New Glenn pad explosion May 2026 | |
P/E, dividend yield, 5-yr avg ROE = n/a / not meaningful (SpaceX is GAAP-loss-making; no dividend). The honest read: SpaceX trades at a multiple (~94x→150x sales) that is >4x Nvidia's and dwarfs every space peer. There is no clean comp — the market is pricing it as a category of one.
Traction / unit economics (add): Starlink ~$11.4B rev / 10M subs ≈ ~$95/mo blended vs. the disclosed ARPU ~$81/mo (mix of higher-ARPU enterprise lifts blended). Segment op margin Starlink ~+39% ($4.4B/$11.4B). This is a genuinely good standalone SaaS-like business buried inside a loss-making conglomerate.
Lens 8 · Stock-Price Catalysts (overlay → funding/product events + post-IPO tape)
Pre-IPO "price" moves were funding step-ups (Lens 5). Post-IPO, the tape is 6 days old but already eventful:
- 2026-06-12 IPO: +19% day one ($135 → $160.95); opened $150, intraday high $176.52. Became the 6th-largest US public company on day one.
- 2026-06-16 Cursor (Anysphere) deal: $60B all-Class-A-stock acquisition of the AI-coding startup — first post-IPO move signals an AI pivot (and ~3.4% dilution); stock "surged past analyst consensus" toward a ~$214–227 area. Live ref $214.80 on 2026-06-16 ``.
- Orbital AI data centers: regulatory ask to deploy up to 1M AI satellites — narrative catalyst tying the AI bet to the launch moat.
Looming catalysts (downward-skewed):
- Lock-up cliffs: 180-day lock-up expires 2026-12-08; early release windows open after Q2 earnings (late Jul/Aug) and from ~Jun 30 for early venture funds — "severe downward pressure" risk as paper gains get monetized. Musk's 6.4B shares locked until 2027-06-12.
- First public earnings (Q2 2026, late July): first GAAP scorecard as a public co — the moment the "aspiration vs. fundamentals" gap gets tested.
What the tape reveals: in its first week SPCX trades on narrative (AI pivot, founder vision) far more than fundamentals — exactly Greifeld's point. Implied vol ~129% confirms the market itself doesn't know what it's worth ``.
Phase C — Judge people & books
Lens 9 · Management
- Track record (exceptional, quantified): Musk + Shotwell built the only company to achieve orbital-class reuse, ~82% commercial launch share, a 10M-subscriber LEO network from scratch, and 165 launches/yr. On pure operating execution this is arguably the strongest management story in the market.
- Tenure & skin in the game: Musk = founder/CEO/CTO/chair since 2002, ~42% equity. Shotwell = President/COO, ~20 yrs, now worth >$2B on the IPO. Skin in the game is total.
- Capital-allocation history: historically reinvested every dollar into reusability and Starlink — value-creating. But the first public-equity decision is a $60B all-stock Cursor buy (not debt paydown, not Starship capex) — an aggressive, dilutive, AI-direction bet that will define the new capital-allocation regime. The Feb 2026 xAI absorption folded a ~$1B/month cash-burning asset into the only cash-generative space company — explicitly to fund xAI from Starlink's cash flows. This is the crux: great operator, but is he allocating the public's capital to a loss-making AI war he might not win?
- Red flags (governance, not accounting): dual-class lets Musk control 82% of votes with 42% of equity and makes him un-removable; non-independent board; "Corporate Opportunities" clause sanctioning self-dealing across Tesla/Boring/Neuralink; key-man risk is extreme and structurally entrenched. Musk's attention is split across SpaceX, Tesla, xAI, X, Neuralink, Boring.
- Archetype: Founder-visionary-controller, taken to its absolute limit. Implies: outstanding for moonshot execution and long-horizon bets; poor for minority-shareholder protections and governance discipline. You are a passenger on Musk's ship — by design.
Lens 10 · Forensic Red Flags
No filings ingested (filings/ empty), so this is web-only + the pre-fetched regulatory file. Areas a forensic analyst would press:
- Segment-reporting / consolidation opacity: the Feb 2026 xAI merger used a triangular structure keeping xAI a wholly-owned subsidiary "to insulate SpaceX from potential legal headaches" — legitimate, but it complicates segment transparency and related-party lines. Watch how the first 10-Q draws the AI-segment boundary.
- Revenue recognition (Starlink): subscription rev is clean, but ARPU falling 18% while subs grow means revenue growth is unit-driven, not price-driven — monitor for any pull-forward or aggressive deferred-revenue treatment in the first public quarter.
- Cash flow vs. earnings divergence: −$4.9B net loss with Starlink throwing off operating cash and Starship/xAI burning it — the consolidated FCF is the number that matters and is not yet sourced (
balance/cashflow rows in financials.csv are blank by design — the "fill-from-S-1" exercise hasn't run). n/a until the S-1 cash-flow statement is ingested.
- Stock-based comp / dilution: $60B Cursor deal is all-stock (~3.4% dilution); using equity as acquisition currency at a ~110x-sales multiple is value-transfer-sensitive. SBC magnitude on a 22,000-person workforce is material and not yet quantified —
n/a.
- Goodwill/intangibles: the xAI ($250B) and Cursor ($60B) deals will load the balance sheet with enormous goodwill — impairment risk if the AI bets sour. First post-deal balance sheet will be the tell.
Regulatory findings (required sub-section):
- SEC (from
regulatory/regulatory-findings.md, fetched 2026-06-18, period 2021-06-18→2026-06-18): 0 Litigation Releases, 0 AAERs naming SpaceX ``. (Note: the file pre-dates the SEC registration; as a brand-new public filer there is no enforcement history yet — expected.)
- Non-SEC enforcement ``:
- FAA: proposed $633,009 in civil penalties for license violations during two 2023 launches; Musk threatened to sue the FAA for "regulatory overreach".
- NLRB: long-running unfair-labor-practice fight — 5th Circuit (Aug 19, 2025) ruled the NLRB's structure likely unconstitutional; NLRB dismissed its complaint against SpaceX on Feb 9, 2026 for lack of jurisdiction (after the NMB ruled SpaceX falls under the Railway Labor Act because "space transport includes air travel"). Net: SpaceX won, but the labor-relations posture is adversarial and politically charged.
- No material FTC/DOJ/FCC monetary settlements surfaced in the search window.
- Item 3 (Legal Proceedings): not available — no 10-K on disk;
n/a until the S-1's legal-proceedings section is ingested.
Summary: No accounting-fraud signals (and no history to find — it just listed). The real "red flags" are governance (entrenched control, self-dealing clause) and forward forensic risk (goodwill from $310B of AI deals, undisclosed consolidated FCF/SBC). Verified via SEC EDGAR EFTS (LR, AAER) + web search + (S-1 Item 3 pending ingest) as of 2026-06-18.
Phase D — Project & stress-test
Lens 11 · IPO-Readiness & Path-to-Tradeable → now Path-as-a-Public-Stock (overlay swap)
The +private "path to tradeable" lens has already resolved — SpaceX is public as of 2026-06-12. So this lens pivots to the forward revenue/value path and the realistic 3-year trajectory.
Forward revenue scenarios (every input labeled):
- Base: FY25 $18.7B → Morgan Stanley ~$160B by 2028 and ~$330B by 2030. That implies a ~3-yr (FY25→FY28) revenue CAGR of ~105%/yr — heroic, driven by Starlink scale + Starship commercialization + AI ($190B AI rev by 2030 in MS's model).
- Bull: Musk's own $1T revenue by 2030 — i.e. ~3x the MS number; requires Starship economics + orbital data centers + AI all compounding.
- Bear: Starlink ARPU keeps falling, residential saturates, Starship slips (it's already behind on HLS), AI bleeds → revenue grows but margins stay negative; revenue lands well under MS's $160B 2028 — call it **$60–90B by 2028 **, with GAAP losses persisting into 2028.
EPS / profitability: SpaceX is GAAP-loss-making (−$4.9B FY25) and gives no near-term profit guide. Building a credible 3-yr EPS path is not supportable with public data — the share count is in flux (IPO + $60B all-stock Cursor + future deals) and consolidated cash-flow isn't disclosed. n/a for a defensible EPS line. The honest forecast is on revenue and segment profitability, not EPS.
The forecast that actually matters (binary, scoreable): Does Starlink's profit + the IPO's ~$75B raise fund the Starship + xAI burn long enough to reach the next value-inflection (Starship operational HLS / orbital-compute revenue) before the lock-up unlocks force a re-rating?
Per --watchlist unattended rules, not logging a forecast.ts Brier forecast (that's reserved for genuinely committed base cases in attended runs). If logged later, the scoreable line would be: "SPCX FY26 revenue ≥ $28B" (Starlink ~$24B + launch ~$4.5B + partial AI) — p≈0.55.
Lens 12 · Bull vs Bear
Bull case. SpaceX is a once-in-a-generation infrastructure monopoly. Reusability gives it a cost moat no one has matched in 20 years of trying; it launches ~82% of the world's commercial payloads and its competitors' satellites. Starlink is a proven, +48%-growth, $4.4B-operating-profit network with 10M subs and a fresh higher-ARPU growth vector in enterprise/aviation/maritime/Direct-to-Cell. Starship unlocks an order-of-magnitude cost step-change and Mars optionality. The AI arm (xAI/Grok + Cursor) plus orbital solar-powered data centers is genuine blue-sky optionality that uses the launch moat as the enabler. Morgan Stanley sees ~$330B revenue by 2030 and $3.4T by 2040. Founder-led, total alignment, flawless operating record. If you believe in the vision, this is the purest way to own the off-Earth economy.
Bear case (permanent-impairment risks).
- Valuation is the risk. ~94x→150x sales (>4x Nvidia) on a company losing $4.9B/yr. Morningstar fair value = $780B — less than half the IPO and ~3.5x below the ~$2.8T live mark; their DCF puts core launch+Starlink at ~$611B EV + only $170B probability-weighted for AI. The market is paying ~$2T+ for a $780B-fair-value business by a credible bear's math.
- AI is a value-destruction vector. Morningstar flags xAI as a "material threat of value destruction," moat "indeterminate". The first public capital-allocation move is a $60B all-stock Cursor bet — diluting holders to fund an AI war against OpenAI/Anthropic it may lose. −$2.5B AI op loss already.
- Key-man + governance. Musk controls 82% of votes, is un-removable, runs 5+ companies, and has a board-sanctioned self-dealing clause. Any Musk shock (health, attention, political/legal) is an un-hedgeable structural risk.
Pre-mortem (18 months out, thesis broke). Most likely failure path: the Dec-2026 lock-up + early-window unlocks flood the market with insider/VC supply into a stock that ran on first-week narrative; Q2/Q3 earnings show widening consolidated losses (AI burn + Cursor dilution + Starship capex) with no GAAP-profit date; Starlink ARPU keeps sliding as Kuiper/Leo undercuts internationally; Starship's HLS slips again. The multiple compresses from ~110x toward Morningstar's ~40x-equivalent fair value — a 50%+ de-rate even with revenue still growing. The business is fine; the price breaks.
Are multiples too high? Yes, on any conventional metric. The only defense is that there is no conventional comp and you're buying a 10-year option on the space+AI economy. That's an aspiration trade, not a valuation trade — and the bull (Greifeld) and the bear (Morningstar) agree on that; they just disagree on whether to pay for it.
Contrarian view (what the market refuses to see): The crowd is debating "is $2T too much for SpaceX?" The thing being under-discussed: Starlink alone, as a standalone, is a cleaner and arguably more valuable asset than the conglomerate — a profitable, +48%-growth global broadband utility. The xAI/Cursor AI bolt-on lowers the quality of the equity by stapling a −$2.5B-and-growing cash incinerator with an "indeterminate moat" to the crown jewel, and funding it with the jewel's cash. The contrarian trade isn't long or short SPCX — it's "the sum is worth less than the Starlink part," and the bull thesis quietly depends on the market never being able to separate them.
Lens 13 · Devil's Advocate (short-seller)
Dismantling the bull case:
- Revenue concentration: 61% of revenue + ~100% of profit is one segment (Starlink), and that segment's ARPU is already falling 18%. The "diversified conglomerate" is really one profitable business subsidizing two loss-makers. If Starlink growth slows (saturation + Kuiper price war), the whole edifice's profit engine stalls.
- Most dangerous competitor bulls underestimate: not Blue Origin (New Glenn just exploded on the pad, May 2026 ) — it's Amazon Leo/Kuiper on the connectivity side (Amazon's balance sheet + AWS bundling can undercut Starlink internationally) and, on launch, the cadence of Rocket Lab's Neutron + Relativity's Terran R arriving 2026–27. "The market only needs two or three winners" — but if even one credible reusable rival reaches cadence, SpaceX's ~82% pricing power erodes at the margin.
- Worst capital-allocation moves: folding a $1B/month-burning xAI into the only cash-generative space company, then spending $60B of public stock on Cursor as the first post-IPO act. A skeptic reads this as the founder using minority shareholders' newly-public currency to bankroll a personal AI ambition — enabled by a self-dealing "Corporate Opportunities" clause and 82% voting control.
- Assumptions that must hold for today's price: Starship works and is cheap and on time (it's behind on HLS as of Mar 2026 ); Starlink reaccelerates at higher ARPU; AI segment turns from −$2.5B to a $190B-by-2030 business; no Musk shock. Strip any one and the ~110x multiple is indefensible.
- If growth disappoints 20–30%: at 94–150x sales, a 25% revenue miss vs. the MS path doesn't just dent EPS (there is no E) — it collapses the narrative, and narrative is the entire valuation support (per Greifeld). A re-rate toward Morningstar's $780B fair value is −55%+ from the ~$2.8T live mark.
- Single scenario that permanently impairs: a Starship catastrophic failure with loss of cadence + a simultaneous Kuiper-led Starlink price war, leaving the launch moat questioned exactly as the AI burn peaks and the lock-up unlocks — forcing a dilutive down-round-equivalent equity raise at a broken multiple. Plausibility: low-to-moderate individually, but the lock-up timing (Dec 2026) makes a sentiment impairment materially likely even without an operational one.
Lens 14 · Management Questions (ordered by information value)
- What is the consolidated free-cash-flow trajectory, and at what date do you expect GAAP profitability for the whole company (not just Starlink)?
- The first use of public equity was a $60B all-stock Cursor acquisition — walk us through the capital-allocation logic versus paying down obligations or funding Starship. What's the hurdle rate?
- xAI lost $2.5B and Morningstar calls its moat "indeterminate." What is the concrete plan and timeline for the AI segment to reach segment-operating breakeven, and how much more capital will it consume?
- Starlink ARPU fell 18% (2023→2025). Where does blended ARPU stabilize, and how much of future growth is units vs. price vs. higher-ARPU verticals (enterprise/aviation/maritime/D2C)?
- Starship's HLS milestones are behind schedule (propellant-transfer + cert review slipped past summer 2025). What is the realistic critical-path date for an operational Starship, and what does it cost to get there?
- How do you think about minority-shareholder protections given 82% voting control, a non-independent board, and the "Corporate Opportunities" self-dealing clause?
- With lock-up early-release windows opening from ~June 30 and the cliff on Dec 8, 2026, how are you managing the supply overhang and insider/VC selling?
- What share of launch capacity is consumed internally by Starlink vs. sold externally, and how do you grow third-party launch revenue beyond +8%?
- Orbital AI data centers / 1M-satellite ask — what's the realistic revenue model, timeline, and capex, and is this funded by Starlink cash or further equity?
- How exposed is Starlink to Amazon Leo/Kuiper internationally on price, and what's your defense if they bundle connectivity with AWS?
- Quantify stock-based compensation and expected dilution over the next three years given an acquisition strategy paid in equity.
- What's the goodwill/intangibles carrying value from the xAI and Cursor deals, and what triggers an impairment test?
- How do you allocate Musk's time and attention across SpaceX, Tesla, xAI, X, Neuralink, and Boring — and what's the succession plan if he's incapacitated?
- What is your exposure to U.S. government contract concentration (NASA/DoD) and the political risk of the founder's public posture toward regulators (FAA/NLRB)?
- How should investors value the company when the CEO says $1T revenue by 2030 and your lead sell-side analyst (Morgan Stanley) says $330B — which assumptions explain the 3x gap?