SpaceX (SPCX): Sum-of-the-Parts — What Investors Are Paying For
Post the February 2026 all-stock absorption of xAI, SPCX is a three-engine conglomerate, not a rocket company: Starlink (the only profit center), Launch (Falcon + Starship), and AI (Grok / X / datacentres). It priced at $135 ($1.77 trillion) on June 12 and traded up sharply on a tiny early float. This note values each part on peer math, locates Starship where it economically lives, and shows the gap between defensible value and the tape.
1. Segment financials (FY2025, from S-1)
| Segment | Revenue | Adj. EBITDA | EBIT |
|---|---|---|---|
| Connectivity (Starlink) 10.3M subs, +86% EBITDA | $11.4B | $7.17B (63%) | +$4.4B |
| Space (Falcon / Starship) $3.0B Starship R&D | $4.09B | $0.65B | −$0.66B |
| AI (xAI / Grok / X) 61% of group capex | $3.2B | neg. | −$6.36B |
| Consolidated | $18.7B | $6.58B | −$2.59B |
Net loss 2025: −$4.9B. Pre-IPO net debt approx. $14B ($30B debt / $16B cash, Q1'26). Operating EV approx. $2.54T post-IPO.
2. Peer-anchored SOTP
| Segment EV | Peer anchor | Bear | Base | Bull |
|---|---|---|---|---|
| Starlink | T-Mobile 10x / Iridium 9-10x; growth-adj 18-22x | $130B | $250B | $380B |
| Space | Rocket Lab approx. 60-90x EV/rev (neg. EBITDA) | $70B | $120B | $220B |
| AI | xAI Feb'26 deal mark approx. $250B; OpenAI $500B | $90B | $200B | $380B |
| Sum (EV) | $290B | $570B | $980B | |
| Equity / share | $24 | $45 | $76 |
Strip out Space ($120B) and AI ($200B) at generous marks, and the residual the market assigns Starlink is approx. $2.22 trillion — roughly 202x 2026E EBITDA ($11B) or 308x 2025 EBITDA, versus T-Mobile 10x and Iridium 9-10x. The entire market cap is an option on Starlink compounding into a global connectivity monopoly several times its current size, plus Starship and AI optionality. None is underwritable from today's segment economics.
3. Where Starship actually lives
Starship is a cost today (the $3.0B R&D that pushes Space to negative EBIT), with 11 non-orbital test flights, first paid payload in H2 2026, and roughly $15B sunk. But its value is not a standalone launch stub — it is the enabler of V3 Starlink satellites (1 Tbps each, 60 per launch, approx. 20x the capacity of a Falcon 9). That capacity unlock drives the entire Starlink forward curve. So Starship belongs inside Connectivity, modelled through Starlink execution paths.
Starlink forward EBITDA ($B) by Starship V3 path
| Year | V3 Success | V3 Delayed | V3 Failure |
|---|---|---|---|
| 2026 | 10.7 | 9.8 | 9.0 |
| 2028 | 22.8 | 15.2 | 11.7 |
| 2030 | 37.9 | 20.5 | 13.5 |
| 2032 | 51.7 | 24.3 | 14.5 |
| Exit mult | 13x | 10x | 7x |
| Starlink EV | $469B | $188B | $93B |
| Probability | 30% | 45% | 25% |
Starship value located inside Connectivity — the V3 delta: Success minus Delayed = $281B; Success minus Failure = $376B. Properly placed, Starship is a roughly $280-380B swing factor inside Starlink, not the $20-70B Space stub. Crediting Starship more does not help the bull case; it merely relocates value. Probability-weighted Starlink EV: $249B.
4. Orbital compute — the bull's best lever
The one pillar the base model omits: in-orbit AI compute (power and compute lofted by Starship, running xAI's models). Modelled aggressively — revenue from 2029 ($2B) to $26B by 2032 at 50% margins, 18x exit — it is worth approx. $137B (about $10/share). Real and material, but it stacks onto Connectivity rather than multiplying it.
| Stacked bull (everything right) | EV |
|---|---|
| Starlink — V3 Success | $469B |
| + Orbital compute | $137B |
| + Falcon (ex-Starship) | $50B |
| + AI / xAI | $200B |
| = Stacked bull EV → approx. $68/share | $857B |
Tape is approx. $196. Even the maximalist stack leaves $1,681B of EV unexplained; the tape's implied Starlink-plus-orbital figure (approx. $2,288B) is 3.8x the everything-right case ($607B).
5. Sensitivity — V3-Success Starlink EV ($B)
| WACC \ exit | 9x | 11x | 13x | 15x | 17x |
|---|---|---|---|---|---|
| 8.0% | 371 | 434 | 496 | 559 | 622 |
| 9.0% | 351 | 410 | 469 | 528 | 587 |
| 11.0% | 315 | 368 | 420 | 473 | 525 |
The entire grid spans $315B–$622B. The tape's implied Starlink number is approx. $2,288B. Even the most favourable corner (8% WACC, 17x exit) is still 3.7x short.
6. The gap survives every lever
| Case | EV | $/share | vs $196 |
|---|---|---|---|
| Probability-weighted SOTP | $499B | $40 | −80% |
| V3-Success (no orbital) | $719B | $57 | −71% |
| Stacked bull (everything right) | $857B | $68 | −65% |
| Damodaran DCF (independent) | $1.22T | approx. $97 | −51% |
| Tape | $2,538B | $196 | — |
Every defensible lever is pulled — Starship working perfectly, Starlink compounding to $76B revenue, a $137B orbital-compute pillar, charitable WACC and exit multiples. The maximalist answer is approx. $68/share against a tape of approx. $196. There is no combination of modelable inputs that bridges to $2.57T; the residual is pure narrative premium on a TAM (AI, Mars) that does not discount to a number.
This is a scarcity / momentum print on a tiny early float with no insider selling yet. Fundamentals re-assert when lockup-adjacent supply arrives.
Assumptions & caveats. WACC 9.0% (Damodaran group approx. 8.4%), D&A 15% of revenue, mid-year discounting, capex tapering 32% to 14% of revenue, NOL-shielded cash tax 5% to 21%. All operating assumptions and segment multiples are analyst interpretation, not S-1 facts; financials are from SpaceX's S-1 / 424B and market data as of mid-June 2026. Shares (approx. 13.1B) are IPO-implied; net cash is approximate pending the first 10-Q. In fairness to the bull: reported terrestrial xAI compute deals, if recurring and margin-positive, could lift the AI segment above $200B (though Colossus economics are likely thin); a 15-year horizon could surface revenue lines no one is modelling. Those are reasons intrinsic value is uncertain, not reasons it is $2.57T today.
Independent research for discussion only. Not investment advice. The author may hold or initiate positions. Do your own diligence.
No redactions applied.
Published via wiis. Originals are encrypted at rest; this page reflects the redacted version. report abuse.
Discussion
No comments yet.