The one-line call
A+ business. C valuation. Monitor — don't chase.
SpaceX owns the two best assets in space — a launch near-monopoly and the only profitable satellite-broadband network on Earth — behind a moat that gets wider every time a rival stumbles. But at a reported ~$800B standalone / ~$2T after the xAI merger, the price already prices in near-flawless execution. For a value investor, there is essentially no margin of safety left.
Read this first
The framework insists I say plainly what I can and can't know, and tier every number: 🟢 High (SEC/S-1, official), 🟡 Medium (credible media), 🔴 Low (estimate, single-source).
- SpaceX was private for most of its life. Nearly all pre-2026 financials are outside estimates, not audited books. Treat the history as directional.
- Two 2026 events change everything (reported, 🟡). SpaceX reportedly acquired xAI (~Feb 2026), folding a loss-making AI business into its accounts, and IPO'd on ~12 June 2026 (Nasdaq: SPCX). Both land near or after my knowledge cutoff, so I flag them as reported — and I keep "space business standalone" separate from "consolidated with xAI" everywhere, because they tell opposite stories.
- Every bull point gets an inverse check. For each strength, I state what would break it.
This is research produced by an AI through a published framework. It is not investment advice.
The verdict in six numbers
Six-dimensional scorecard
| Dimension | Score | Why |
|---|---|---|
| Business quality | ★★★★★ | Launch near-monopoly + only profitable LEO broadband network |
| Economic moat | ★★★★★ | Reusability cost gap, cadence flywheel, spectrum, vertical integration — and widening |
| Management | ★★★★☆ | Elite execution (Shotwell) & bold capital allocation, but severe key-man + political risk |
| Financial strength | ★★★★☆ | Starlink is cash-generative; xAI burns ~$6.4B/yr; capex ~$20B (space ★★★★☆ / consolidated ★★★☆☆) |
| Industry / trend | ★★★★★ | Right side of a $1.8T-by-2035 civilizational S-curve |
| Margin of safety | ★★☆☆☆ | Wonderful business, demanding-to-rich price |
Composite: a wonderful business (5/5 quality and moat) trading at a price that has already discounted the wonder.
1 · What it actually is
The flywheel, in one breath
SpaceX makes rockets cheap by flying the same rocket again and again, then uses that cheap launch to fill the sky with its own internet satellites — and sells the internet (Starlink) to consumers, planes, ships, and militaries.
Three engines:
- Connectivity / Starlink — ~61% of revenue and the profit engine. $11.4B in 2025 (+50%), ~63% segment EBITDA margin, 12M subscribers across 155+ countries. 🟢
- Space / Launch — ~22%. $4.09B in 2025. Sells rides to NASA, the DoD, the NRO, commercial operators — and to itself. 🟢
- AI / xAI — ~17%, new in 2026. $3.2B revenue but a ~$6.4B operating loss. Dilutive to quality and cash. 🟢
The lollapalooza: reusability → cheap launch → SpaceX launches its own Starlink cheaply → Starlink cash funds Starship and more launches → higher cadence lowers cost again. Rivals who buy launch from SpaceX fund the very flywheel trying to bury them.
The right question isn't "what is it worth" — it's "will this business be obviously bigger and better in ten years?" Starlink selling a subscription to anyone with a clear view of the sky, that I understand. xAI bolted on at a $1.25 trillion combined price, I do not. — Duan Yongping (段永平) lens · business essence / 本分
Inverse check. Starlink ARPU keeps falling and never inflects to profit-per-user; Starship's cost-collapse fails so the next-gen satellite economics never arrive; xAI's burn forces the crown-jewel space business to subsidize an AI arms race.
2 · Follow the money
Two companies wearing one ticker
The space business is operating-cash-flow and EBITDA positive. The consolidated entity is GAAP-loss-making — because of ~$20.7B capex and xAI's loss. Same page, opposite stories.
| Consolidated (S-1, 🟢) | 2023 | 2024 | 2025 |
|---|---|---|---|
| Revenue | ~$10.4B | ~$14.0B | $18.67B |
| GAAP net income / (loss) | — | ~+$0.79B* | ($4.94B) |
| Adjusted EBITDA | — | — | +$6.58B |
| Operating cash flow | — | — | ~+$6.8B |
| Capex | — | — | ~$20.7B |
*2024 profit is on the older standalone basis; the 2025 net loss is driven almost entirely by consolidating xAI.
What it's worth — five methods
| Method | Output (space business, ex-xAI) | Tier |
|---|---|---|
| Latest round (adjusted) | ~$800B (Dec 2025 secondary, $421/sh) | 🟡 |
| Comparable multiples | ~$250–450B | 🟡 |
| DCF, 3 scenarios | Bear ~$300B / Base ~$650B / Bull ~$1.3T+ | 🔴 |
| Terminal reverse-engineering | ~$800B implies Starlink captures the entire 2030 sat-internet TAM and Starship cuts $/kg 10× | 🔴 |
| M&A / transaction | No comparable asset trades — there is no second SpaceX | — |
The gap between $250B of fundamentals and ~$800B+ of market is the option value on Starship and defense. That option is real — but unproven.
It's a wonderful business. But you can pay too much for a wonderful business. At thirty to forty times forward sales you're not buying the company — you're buying a flawless future. — Warren Buffett lens · margin of safety
3 · The moat, stress-tested
Every rival stumbled on schedule
The clearest read on the moat in 2025–26 is that the competition kept failing on time.
| Competitor | Status, mid-2026 | Threat |
|---|---|---|
| Blue Origin (New Glenn) | 3 flights; pad explosion 28 May 2026 grounded it | Real long-term, years behind now |
| Rocket Lab (Neutron) | Unflown; slipped to ≥Q4 2026 after tank rupture | Medium-lift threat deferred |
| ULA (Vulcan) | NSSL launches paused Feb 2026 after SRB anomaly | Fading |
| China (Guowang + Qianfan) | 93 launches (record); constellations ~1.4% built; no reusable landing yet | The real 2030s threat |
| Amazon Leo (ex-Kuiper) | ~10% deployed; missed FCC deadline; enterprise beta only | Best-funded rival — 30× behind, buying Falcon 9s to catch up |
Moat sources: reusability & scale ★★★★★ · cadence flywheel ★★★★★ · spectrum/regulatory ★★★★☆ · government entrenchment ★★★★☆ (~$6.45B Space Force awards, May 2026). Trend: widening.
The question is which company sits on the right side of a genuine revolution with a durable advantage. SpaceX plainly does. What worries the long-term owner isn't the moat — it's the price you pay to stand inside it, and whether one man's other ambitions drain it. — Li Lu (李录) lens · the long game
4 · What breaks it
Risk & governance
You would be a minority passenger. Musk holds ~42% and majority voting — capital allocation is his call, proven brilliant and idiosyncratic (the xAI merger is Exhibit A). The saving grace is Gwynne Shotwell, who supplies the operational reliability the genius-founder model usually lacks.
The fatal signals — "thesis broken" triggers
- Musk exits operational control, or a public Musk-vs-administration rupture visibly reroutes DoD/NASA dollars → stop / reassess
- Starship suffers a multi-year stand-down → option value evaporates → fair value resets to ~$250–450B → avoid / reduce
- Starlink cash starts cross-subsidizing xAI's losses at scale → quality impairment → reduce
Show me the incentive and I'll show you the outcome. A controlling founder who just merged his money-losing AI startup into his crown-jewel rocket company at a trillion-dollar mark has told you exactly whose interests set the price. — Charlie Munger lens · invert, always invert
The decision
Who should do what
| If you are… | Call | Why |
|---|---|---|
| A value investor | Monitor / avoid | Wonderful business, wrong price; no margin of safety |
| A pre-IPO secondary holder | Hold, trim to <5% | You own the moat cheaply — let it run, manage concentration |
| A growth / momentum buyer | Speculative only | Justifiable if you underwrite the option value and can stomach a 50%+ drawdown |
| A new public-market entrant | Wait for a de-rate | Buying at ~30–40× forward sales into a hype-priced debut |
What flips me constructive: a 20–35% de-rate that restores a safety margin · or 2–3 fully-successful Starship reuse cycles · or xAI ring-fenced so Starlink cash isn't at risk.
Four-master synthesis: all four salute the business — moat, trend, execution. All four balk at the price and the xAI complication. A+ business, C valuation → monitor, do not chase.
Selected sources
SEC S-1 (financials, 🟢) · CNBC / NPR / Fortune / Bloomberg (valuation & IPO, 🟡) · CNBC / TechCrunch (xAI merger, 🟡) · SpaceX 2025 Progress Report, Quilty Space, planet4589.org (Starlink scale) · SpaceNews / Space.com (launch cadence & competitors) · DefenseScoop / SpaceNews (government awards) · McKinsey/WEF, Morgan Stanley (TAM). Confidence tiers and inverse-risk checks applied per the AI Berkshire private-company-research methodology. Post-cutoff events (xAI merger, IPO, mid-2026 competitor events) are reported via web search, not independently verifiable by the analyst, and are tiered 🟡 accordingly.
Produced with the AI Berkshire framework · styled in Plainspoken (Cool Slate). Research, not investment advice.