SpaceX
Space Exploration Technologies Corp.
The best company we have ever been unable to recommend
Three-Statement Model
SpaceX (Nasdaq: SPCX) — Three-Statement Model
Coverage desk · Tier 1 (SEC-reporting, listed 2026-06-12) · As of 2026-07-15. Author: Sonnet
(model-build stage), per PROCESS.md §3.3/§5. Companion to filings.md, research.md.
Re-tier note. This model was scoped under the desk's Tier 2 (private) roster entry but is
built Tier 1: FY23–FY25 actuals are anchored to SpaceX's S-1 (filed 2026-05-20, amended
2026-06-01/03) and 424B4 (2026-06-12), corroborated by Q1 2026 disclosure. No 10-K/10-Q has
been filed yet (first 10-Q — for the quarter ended 2026-06-30 — is not yet out), so FY26E
onward is entirely desk estimate, calibrated to the Q1 2026 print (revenue $4.694B, operating
loss -$1.943B, adjusted EBITDA +$1.127B, capex ~$10.1B of which ~$7.7B AI infra — S-1/FWP via
filings.md §2b, 2026-05-20 vintage). Segment naming follows the S-1's own three-segment
structure: Space, Connectivity (Starlink), AI. Where the S-1 does not break out a
sub-line (e.g., external-vs-internal launch mix, capex-by-category), figures are desk estimates
labeled E with rationale stated. Shares outstanding used throughout (~13.0B) are
back-solved from $1.75T ÷ $135 IPO price — not a disclosed figure; flagged in filings.md §2c
as pending confirmation against the 424B4 cover page. $ in millions unless noted.
1. Assumption stack
Revenue is built bottoms-up per segment driver, not as a top-line growth rate, per PROCESS.md
§5. A critical structural point drives the Space segment: Falcon 9/Heavy launch cadence
(165 flights in 2025) is dominated by SpaceX's own internal Starlink deployment missions, which
generate no external launch-services revenue — only externally contracted (commercial +
government) launches are revenue-bearing. This is why Space-segment revenue ($4.1B actual,
FY25) is far smaller than cadence × list price would suggest.
1a. Space (external launch + government programs + crew/cargo)
| Driver | FY23 | FY24 | FY25 | FY26E | FY27E | FY28E | FY29E | FY30E | Rationale |
|---|---|---|---|---|---|---|---|---|---|
| External (revenue-bearing) launches, # | 30E | 34E | 35E | 38E | 40E | 42E | 44E | 46E | Total Falcon cadence 96→134→165→~140-145E (2023-25 actual, 2026E per Shotwell guide, Time/Octagon AI, 2026); desk estimates ~65-80% of cadence is Starlink-internal, leaving external launches roughly flat-to-modestly-up even as total cadence plateaus, with Starship dedicated missions adding external volume from FY28E |
| Avg. price / external launch ($M) | $65E | $68E | $71E | $76E | $80E | $85E | $88E | $90E | Falcon 9 dedicated list $67M (early-2026)→$74M (late-Feb-2026) — SatBase, 2026; blended toward the disclosed $74M print by FY26E and toward Starship's $90M contracted price (Voyager Technologies 10-K, cited Motley Fool 2026-03-21) as Starship missions layer in from FY28E |
| → External launch revenue | 1,950 | 2,310 | 2,490 | 2,888 | 3,200 | 3,570 | 3,872 | 4,140 | Product of above |
| Government programs & services (NASA HLS/CCtCap/CRS milestones, Starshield, NSSL Lane 1/2, Golden Dome) | 4,000E | 3,340E | 890E | 1,312E | 2,100E | 2,930E | 4,028E | 5,160E | Residual to match disclosed/estimated Space-segment totals; lumpy by milestone-recognition timing. Backlog context: NASA HLS ~$4.04B through 2028, CCtCap $4.93B total, CRS-1+2 $4.8B (NASA/SpaceNews); NSSL Phase 3 Lane 2 ~$5.9B/28 missions 2027-32 (Spaceflight Now, 2025-04-05); Golden Dome ~$6.45B awarded within 4 days, May 2026 (Bloomberg, 2026-05-29); company-cited ~$22B total federal backlog. FY25 dip reflects Starship R&D absorption into segment cost, not revenue loss — segment is loss-making on that basis per S-1 |
| Memo: Crew/cargo/extended missions (private astronaut + NASA Dragon flights) | 600E | 650E | 700E | 800E | 900E | 1,000E | 1,100E | 1,200E | Additive component of Space segment revenue below (verified: External launch revenue + Government programs + this line foots exactly to the segment total every year, e.g. FY25 2,490+890+700=4,080) — not embedded within the Government-programs line above it, despite the "memo" label; shown as a separate driver row for transparency into launch-adjacent recurring revenue. NASA CCtCap/CRS cadence (~4-6 Dragon flights/yr) plus a growing handful of private-astronaut missions (Axiom/Polaris/Fram2 precedent) at ~$55-70M/seat-equivalent |
| Space segment revenue | 6,550E | 6,300E | 4,080A | 5,000E | 6,200E | 7,500E | 9,000E | 10,500E | FY25 = S-1 actual ($4.1B, HL.co.uk S-1 breakdown, 2026); FY23-24 back-solved from S-1-disclosed Connectivity revenue-share mix; FY26E+ grows on Golden Dome/NSSL Lane 2 ramp + Starship-enabled external launch growth |
1b. Connectivity (Starlink)
| Driver | FY23 | FY24 | FY25 | FY26E | FY27E | FY28E | FY29E | FY30E | Rationale |
|---|---|---|---|---|---|---|---|---|---|
| Active subscribers, avg during year (M) | 1.65E | 3.45A | 7.05E | 15.75E | 27.0E | 36.0E | 43.5E | 50.5E | 1.0M (Dec-22) → 2.3M (end-23) → 4.6M (end-24) → ~9.5M (end-25E) → 12M+ (confirmed 2026-06-04) → mgmt target >25M by YE26 (SDxCentral); desk haircuts mgmt's 25M target to ~22M year-end (implying ~15.75M average) and grows subscriber adds at a decelerating rate into the high-40s/low-50s M by FY30E as the reachable satellite-only population (~650-750M per Novaspace) gets penetrated |
| Blended ARPU ($/mo) | $99A | $91A | $81A | $62E | $58E | $55E | $53E | $52E | $99 (2023) → $91 (2024 avg) → $81 (2025 avg) → $66 (Q1'26, actual) — The Information/Roic.ai, 2026-04-29; S-1 via TheStreet. Desk continues the disclosed volume-over-price strategy (emerging-market mix shift) at a decelerating rate of decline, stabilizing near $52/mo as enterprise/D2C mix offsets consumer dilution |
| → Core consumer subscription revenue | 1,960 | 3,767 | 6,853 | 11,718 | 18,792 | 23,760 | 27,666 | 31,512 | Avg subs × ARPU × 12 |
| Non-subscription uplift (hardware/kit sales, business/maritime/aviation, Direct-to-Cell wholesale, Starshield/gov broadband) | 1,890E | 3,933E | 4,537E | 4,782E | 2,708E | 2,740E | 3,834E | 4,988E | Residual to segment total; large early (heavy $500-600 hardware-kit sales per gross add against a small base, lumpy enterprise/Starshield wins), shrinking to a steady-state ~12-16% of core as the subscriber base matures and per-unit hardware revenue becomes small relative to the recurring base |
| Connectivity segment revenue | 3,850E | 7,700A | 11,390A | 16,500E | 21,500E | 26,500E | 31,500E | 36,500E | FY25 = S-1 actual $11.39B, +48% YoY off FY24's $7.7B (Via Satellite S-1 coverage, 2026-05-20); FY23 back-solved from S-1's disclosed 37% FY23 Connectivity revenue share (HL.co.uk); FY26E+ decelerating from +45% (FY26E) to +16% YoY (FY30E) as subscriber growth outpaces further ARPU decline, consistent with the FY25 pattern |
1c. AI (SpaceXAI / xAI / Grok / X)
New segment; formalized by the 2026-02-02/03 all-stock SpaceX-xAI merger (implied combined
value $1.25T — CNBC, 2026-02-03), disclosed for the first time in the S-1 as a $3.2B FY25
revenue line. Note excluded from modeling: a single-sourced, unverified claim of an
Anthropic cloud-compute agreement ($1.25B/mo through May 2029) circulating in secondary S-1
commentary is not used as a driver input per the disclosure sweep's explicit flag
(financials.md §1) — it has not been corroborated against primary S-1 text or Anthropic's own
disclosures.
| Driver | FY23 | FY24 | FY25 | FY26E | FY27E | FY28E | FY29E | FY30E | Rationale |
|---|---|---|---|---|---|---|---|---|---|
| Grok subscriptions + API revenue | — | — | 1,700E | 2,900E | 4,550E | 6,800E | 9,585E | 12,775E | Segment pre-existed the formal Feb-2026 merger only immaterially; FY25 total is S-1-disclosed. Desk assumes Grok/API mix rises from ~53% (FY25) to ~73% (FY30E) of segment revenue as compute/API monetization scales faster than advertising — no direct comp disclosed, a judgment call flagged as the single largest source of forecast uncertainty in this model |
| X advertising revenue | — | — | 1,500E | 1,900E | 2,450E | 3,200E | 3,915E | 4,725E | X folded into xAI Mar-2025 per disclosure sweep; ad revenue assumed to grow modestly off a depressed base, well below the subscription/API line's growth rate |
| AI segment revenue | 0 | 0 | 3,200A | 4,800E | 7,000E | 10,000E | 13,500E | 17,500E | FY25 = S-1 actual (Via Satellite/HL.co.uk/market.md, 2026-05-20); Q1'26 run-rate $818M/qtr (~$3.3B annualized, KraneShares/Vested Finance S-1 takeaways) supports the FY26E base but implies acceleration through the year is needed to hit $4.8B — the single most load-bearing, least-certain line in the model. AI segment op. loss was -$6.355B on $3.2B revenue in FY25 alone (exceeds total company net loss) — profitability here, not revenue growth, is the real swing factor (see §5) |
Total revenue check ($M)
| FY | Space | Connectivity | AI | Total | Source |
|---|---|---|---|---|---|
| FY23 | 6,550E | 3,850E | 0 | 10,400A | S-1 total actual |
| FY24 | 6,300E | 7,700A | 0 | 14,000A | S-1 total actual |
| FY25 | 4,080A | 11,390A | 3,200A | 18,670A | S-1 total actual |
| FY26E | 5,000E | 16,500E | 4,800E | 26,300E | |
| FY27E | 6,200E | 21,500E | 7,000E | 34,700E | |
| FY28E | 7,500E | 26,500E | 10,000E | 44,000E | |
| FY29E | 9,000E | 31,500E | 13,500E | 54,000E | |
| FY30E | 10,500E | 36,500E | 17,500E | 64,500E |
2. Income statement, FY23–FY30E ($M)
Margin assumptions: blended gross margin rises from 38% (FY23) toward 60% (FY30E) as Connectivity's
share of the mix grows and AI compute utilization improves at scale — but the path is not smooth:
FY24-25 COGS is a desk reconstruction (not an S-1 line item) forced to satisfy the
GP − R&D − SG&A = adjusted-EBITDA identity against the S-1-disclosed EBITDA actuals (see note
below the table), which pushes reconstructed FY24/FY25 gross margin to 54.1%/74.6% — a spike
above the FY30E target, not a smooth ramp. This is flagged as a desk-model audit finding, not
resolved further here (see model-audit note below); it implies the true FY24-25 COGS/R&D/SG&A
split, once the S-1's actual line-item income statement is pulled, will look materially different
from what is shown. R&D is Starship-flagged
(the S-1 states FY25 Space-segment R&D of $3.0B "the entirety of which went to the Starship
program" — BNN Bloomberg via financials.md); AI/other R&D scales with Grok model development.
Stock-based compensation (SBC) is broken out as a non-operating add-back item, consistent with
the gap between S-1-disclosed adjusted EBITDA (excl. SBC) and GAAP net income — the desk's
read of why FY25 swung from FY24's rare GAAP profit to a $4.9B loss despite +33% revenue growth
and positive adjusted EBITDA (financials.md §2).
| ($M) | FY23A | FY24A | FY25A | FY26E | FY27E | FY28E | FY29E | FY30E |
|---|---|---|---|---|---|---|---|---|
| Space | 6,550 | 6,300 | 4,080 | 5,000 | 6,200 | 7,500 | 9,000 | 10,500 |
| Connectivity (Starlink) | 3,850 | 7,700 | 11,390 | 16,500 | 21,500 | 26,500 | 31,500 | 36,500 |
| AI | 0 | 0 | 3,200 | 4,800 | 7,000 | 10,000 | 13,500 | 17,500 |
| Total revenue | 10,400 | 14,000 | 18,670 | 26,300 | 34,700 | 44,000 | 54,000 | 64,500 |
| YoY % | — | +35% | +33% | +41% | +32% | +27% | +23% | +19% |
| COGS | 6,448 | 6,420 | 4,740 | 11,080 | 14,636 | 18,320 | 22,020 | 25,662 |
| Gross profit | 3,952 | 7,580 | 13,930 | 15,220 | 20,064 | 25,680 | 31,980 | 38,838 |
| Gross margin % | 38.0% | 54.1% | 74.6% | 57.9% | 57.8% | 58.4% | 59.2% | 60.2% |
| R&D (of which Starship) | 3,000 (2,200) | 2,500 (1,900) | 4,500 (3,000) | 6,000 (3,300) | 7,200 (3,600) | 8,300 (3,900) | 9,300 (4,100) | 10,200 (4,300) |
| SG&A | 1,800 | 2,140 | 2,850 | 3,420 | 4,164 | 5,060 | 5,940 | 6,708 |
| EBITDA (adjusted, pre-SBC) | (848) | 2,940 | 6,580 | 5,800 | 8,700 | 12,320 | 16,740 | 21,930 |
| EBITDA margin % | (8.2%) | 21.0% | 35.2% | 22.1% | 25.1% | 28.0% | 31.0% | 34.0% |
| D&A | 2,900 | 2,000 | 9,680 | 9,050 | 13,950 | 17,550 | 19,800 | 20,900 |
| EBIT | (3,748) | 940 | (3,100) | (3,250) | (5,250) | (5,230) | (3,060) | 1,030 |
| Interest income/(expense), net | 50 | 150 | 200 | 1,800 | 900 | 400 | 200 | 250 |
| Stock-based comp & other non-operating (net) | (890) | (300) | (2,000) | (1,600) | (1,850) | (2,200) | (2,550) | (2,850) |
| Pretax income | (4,588) | 790 | (4,900) | (3,050) | (6,200) | (7,030) | (5,410) | (1,570) |
| Tax | 12 | 0 | 0 | 150 | 200 | 250 | 300 | 350 |
| Net income | (4,600) | 790 | (4,900) | (3,200) | (6,400) | (7,280) | (5,710) | (1,920) |
FY23A/FY24A/FY25A net income and FY25 adjusted EBITDA tie to S-1-disclosed actuals (-$4.6B,
+$0.791B [rounds to 790 here], -$4.9B, $6.58B respectively — financials.md §2). COGS/R&D/SG&A/D&A splits
for FY23-25 are desk reconstructions calibrated to those four hard numbers, since the S-1's
full line-item income statement was not independently pulled in this pass (flagged in
filings.md §6 as an open item). Model-audit correction (2026-07-15): the FY24 and FY25
COGS figures as originally drafted (7,560 and 6,040) did not satisfy the model's own
GP − R&D − SG&A = adjusted-EBITDA identity — they implied EBITDA of 1,800 (FY24) and 5,280
(FY25) against the stated/actual EBITDA of 2,940 and 6,580, a $1.14B and $1.30B internal
inconsistency respectively. Corrected by solving COGS for consistency (R&D and SG&A held fixed,
since FY25 R&D carries the sourced $3.0B Starship figure and SG&A is unsourced but was not the
locus of the error) to 6,420 (FY24) and 4,740 (FY25); EBITDA, EBIT, pretax, net income, and the
balance-sheet/cash-flow ties (which run off NI and D&A, not COGS) are unaffected by this fix. The
resulting FY24/25 gross margin (54.1%/74.6%) is not a smooth ramp and sits above the FY30E
target — see the flag in §2 margin-assumptions text above; this points to the underlying
COGS/R&D/SG&A bucketing for FY24-25 needing to be re-derived from actual S-1 line items rather
than back-solved, which the desk should treat as an open item, not a resolved one. Read-through for FY26E-30E: under the base case, GAAP
losses widen through FY27E-28E (AI-driven D&A scaling faster than EBITDA) before narrowing
sharply and nearly reaching GAAP EBIT breakeven by FY30E — the AI capex/depreciation cycle, not
the core launch/Starlink business, is what keeps this company GAAP-loss-making through most of
the forecast window despite EBITDA scaling to $21.9B (34% margin) by FY30E.
3. Balance sheet, FY23–FY30E ($M)
Depreciation policy (stated assumption): Starlink satellites depreciated straight-line over 5 years (short design life is industry-standard for LEO constellations); launch/ground infrastructure over 15 years; AI/compute hardware over 4 years (GPU-cluster convention). Blended rate rises from ~18% (FY23-24) toward ~24% (FY26E) as AI capex — which depreciates fastest — becomes the majority of the capex mix, tapering to ~20% by FY30E as that mix stabilizes.
No pre-IPO balance sheet was ever publicly disclosed (Tier 2 for SpaceX's entire history through 2025). The FY22 year-end (opening) balance sheet below is therefore a fully-estimated, internally-consistent starting point (all lines flagged E), constructed to balance and then rolled forward using the S-1-disclosed income-statement actuals (revenue, net income, adjusted EBITDA) for FY23-25. "Other liabilities" is used as the explicit balancing/plug line each year (accrued compute-capacity commitments, accrued capex payables, and other unmodeled items are the plausible real-world composition) — this is stated, not hidden, and the check row below confirms the balance sheet ties exactly every year by construction.
| ($M) | FY22 (opening, E) | FY23A/E | FY24A/E | FY25A/E | FY26E | FY27E | FY28E | FY29E | FY30E |
|---|---|---|---|---|---|---|---|---|---|
| Cash | 8,000 | 5,840 | 2,430 | 3,010 | 38,010 | 16,410 | 2,080 | 2,120 | 2,050 |
| Receivables | 320 | 730 | 980 | 1,400 | 1,970 | 2,600 | 3,300 | 4,050 | 4,840 |
| Inventory | 400 | 967 | 1,134 | 906 | 1,662 | 2,195 | 2,748 | 3,303 | 3,849 |
| PP&E, net (satellites + launch infra) | 16,000 | 17,500 | 26,700 | 37,720 | 60,670 | 79,720 | 94,170 | 104,370 | 112,470 |
| Other assets | 230 | 520 | 700 | 933 | 1,315 | 1,735 | 2,200 | 2,700 | 3,225 |
| Total assets | 24,950 | 25,557 | 31,944 | 43,969 | 103,627 | 102,660 | 104,498 | 116,543 | 126,434 |
| Payables | 300 | 1,290 | 1,512 | 1,208 | 2,216 | 2,927 | 3,664 | 4,404 | 5,132 |
| Deferred revenue (Starlink prepays) | 200 | 578 | 1,155 | 2,050 | 2,970 | 4,300 | 5,300 | 6,930 | 8,030 |
| Debt | 1,500 | 3,500 | 7,000 | 19,500 | 4,500 | 4,500 | 7,500 | 18,500 | 23,000 |
| Other liabilities (plug) | 230 | 429 | 1,427 | 3,261 | 4,841 | 6,383 | 8,564 | 10,399 | 13,032 |
| Total liabilities | 2,230 | 5,797 | 11,094 | 26,019 | 14,527 | 18,110 | 25,028 | 40,233 | 49,194 |
| Paid-in capital | 27,720 | 29,360 | 29,660 | 31,660 | 106,010 | 107,860 | 110,060 | 112,610 | 115,460 |
| Retained earnings (deficit) | (5,000) | (9,600) | (8,810) | (13,710) | (16,910) | (23,310) | (30,590) | (36,300) | (38,220) |
| Total equity | 22,720 | 19,760 | 20,850 | 17,950 | 89,100 | 84,550 | 79,470 | 76,310 | 77,240 |
| Check: Assets − (Liab + Equity) | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
Notes: (1) Debt was historically minimal per the disclosure sweep, but the model shows a
material FY24-25 build (to $19.5B) — this reflects the desk's read that capex/revenue reaching
111% in FY25, with no primary equity raised between Jan-2023 and the June-2026 IPO
(financials.md §5), forced debt-funded bridge financing ahead of the listing; roughly $15B of
that debt is assumed repaid at IPO close from proceeds, consistent with typical S-1 "Use of
Proceeds" language (unconfirmed against the actual 424B4 text — flagged in filings.md §2d).
(2) Paid-in capital jumps from ~$31.7B to ~$106.0B in FY26E on the ~$75B gross IPO raise (net of
~$2.25B assumed underwriting fees) plus ongoing SBC credits — consistent with the June 2026 IPO
being the first primary capital event since January 2023. (3) The FY28E-30E return to debt
issuance (to $23.0B by FY30E) is the balance-sheet expression of the same "self-funding stress"
flagged across all four research notes: even post-$75B-IPO, sustained ~$29-33B/yr capex outruns
operating cash flow in the base case, requiring recurring external financing.
4. Cash flow, FY23–FY30E ($M)
Capex is split Starship/launch infra vs. Starlink constellation vs. AI infrastructure. The
FY25 AI-infra capex figure ($12.7B of $20.7B total) is a hard, sourced data point (Yahoo
Finance/AOL citing SpaceX disclosure, via dynamics.md); FY23-24 and FY26E+ splits are desk
estimates. CFF excludes all pre-IPO tender/secondary transactions (Dec-2024 $350B mark, Jul-2025
$400B mark, Dec-2025 $800B mark) — these were non-dilutive liquidity events between existing
shareholders, not capital raised by the company, per the explicit modeling instruction and the
disclosure sweep's own framing (disclosure.md §2-3).
| ($M) | FY23A/E | FY24A/E | FY25A/E | FY26E | FY27E | FY28E | FY29E | FY30E |
|---|---|---|---|---|---|---|---|---|
| Net income | (4,600) | 790 | (4,900) | (3,200) | (6,400) | (7,280) | (5,710) | (1,920) |
| + D&A | 2,900 | 2,000 | 9,680 | 9,050 | 13,950 | 17,550 | 19,800 | 20,900 |
| + Stock-based comp (non-cash) | 890 | 300 | 2,000 | 1,600 | 1,850 | 2,200 | 2,550 | 2,850 |
| +/- Change in working capital | 300 | 1,200 | 2,000 | 1,800 | 2,000 | 2,200 | 2,400 | 2,600 |
| CFO | (510) | 4,290 | 8,780 | 9,250 | 11,400 | 14,670 | 19,040 | 24,430 |
| Capex — Starship/launch infra | (2,000) | (3,200) | (3,000) | (4,000) | (4,200) | (4,400) | (4,500) | (4,600) |
| Capex — Starlink constellation | (2,400) | (8,000) | (5,000) | (6,000) | (6,800) | (7,200) | (7,500) | (7,900) |
| Capex — AI infrastructure | 0 | 0 | (12,700) | (22,000) | (22,000) | (20,400) | (18,000) | (16,500) |
| CFI (total capex) | (4,400) | (11,200) | (20,700) | (32,000) | (33,000) | (32,000) | (30,000) | (29,000) |
| Primary equity issuance | 750 | 0 | 0 | 72,750 | 0 | 0 | 0 | 0 |
| Debt issuance/(repayment), net | 2,000 | 3,500 | 12,500 | (15,000) | 0 | 3,000 | 11,000 | 4,500 |
| CFF | 2,750 | 3,500 | 12,500 | 57,750 | 0 | 3,000 | 11,000 | 4,500 |
| Net change in cash | (2,160) | (3,410) | 580 | 35,000* | (21,600) | (14,330) | 40 | (70) |
| Beginning cash | 8,000 | 5,840 | 2,430 | 3,010 | 38,010 | 16,410 | 2,080 | 2,120 |
| Ending cash (CF statement) | 5,840 | 2,430 | 3,010 | 38,010* | 16,410 | 2,080 | 2,120 | 2,050 |
| Ending cash (balance sheet, §3) | 5,840 | 2,430 | 3,010 | 38,010 | 16,410 | 2,080 | 2,120 | 2,050 |
| Tie-out check | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
* FY26E net change of 35,000 reflects CFO 9,250 − CFI 32,000 + CFF 57,750 = 35,000; the $75B
gross IPO raise (CNBC/TechCrunch, 2026-06-12) is the dominant driver, partly offset by ~$2.25B
of assumed underwriting fees and a $15B assumed paydown of pre-IPO bridge debt, both netted into
the CFF line above. Primary equity raises are excluded in every year except FY26E (the IPO) and
the small $750M January-2023 round (actual, CNBC 2023-01-02) — consistent with the January-2023
→ June-2026 primary-capital drought documented in financials.md §5.
5. Scenarios (bear / base / bull)
Five swing assumptions drive the spread, ranked by materiality:
- AI/xAI monetization — the single largest swing factor per every research note. Bear: Grok/API/X-ads growth stalls as OpenAI/Anthropic/Google out-compete on model quality and enterprise distribution; AI segment revenue plateaus near $8B by FY30E, losses persist. Bull: xAI achieves hyperscaler-comparable monetization; AI segment scales past $30B by FY30E and turns EBITDA-positive by FY28E-29E.
- Starlink subscriber growth vs. ARPU — bear: subscriber growth decelerates (Amazon Leo/ Guowang-Qianfan competition bites, D2C cannibalizes premium ARPU) and ARPU keeps falling faster than volume offsets it; bull: management's >25M-by-YE26 trajectory (SDxCentral) continues compounding toward 60M+ by FY30E with ARPU stabilizing on enterprise/D2C mix.
- AI capex intensity / self-funding — bear: AI infra capex stays >70% of total capex with no offsetting revenue, capping EBITDA margin expansion and forcing repeated debt/equity raises; bull: capex growth decelerates in absolute terms after FY27E-28E as compute utilization efficiency improves, easing the self-funding stress flagged in §3-4.
- Starship cadence/reusability — bear: the FAA-grounding pattern (Flights 8, 9, 12 all
mishaps within 15 months —
filings.md§3c) continues, Space segment stays structurally loss-making and Starship-dedicated commercial launches never scale; bull: Starship reaches repeatable reuse, unlocking the sub-$500/kg cost target and a meaningful Starship-dedicated external-launch revenue stream by FY28E-30E. - Government/defense execution — bear: Golden Dome/NSSL Lane 2 awards slip or are re-competed amid federal budget pressure (~20% of FY25 revenue is federal — 24/7 Wall St., 2026-05-21); bull: the ~$22B disclosed federal backlog converts to revenue on schedule and SpaceX captures a disproportionate share of follow-on Golden Dome tranches.
| ($M, FY30E) | Bear | Base | Bull |
|---|---|---|---|
| Connectivity (Starlink) revenue | 24,000 | 36,500 | 46,000 |
| Space revenue | 7,000 | 10,500 | 15,000 |
| AI revenue | 8,000 | 17,500 | 35,000 |
| Total revenue | 39,000 | 64,500 | 96,000 |
| EBITDA margin | 18% | 34% | 42% |
| EBITDA | 7,020 | 21,930 | 40,320 |
6. Valuation bridge
Inputs: shares outstanding ~13.0B (back-solved, $1.75T ÷ $135 IPO price — unconfirmed
pending 424B4 cover page, filings.md §2c); last close $136.12 (2026-07-14, Nasdaq); current
implied market cap ≈ 13.0B × $136.12 ≈ $1,770B; current net cash (post-IPO, mid-2026 estimate
netting ~$75B gross proceeds against ~6 weeks of capex burn and partial bridge-debt paydown) ≈
$50B net cash, giving current EV ≈ $1,720B.
6a. Current pricing — implied multiples (base-case estimates)
| FY26E | FY27E | FY28E | |
|---|---|---|---|
| Revenue | 26,300 | 34,700 | 44,000 |
| EBITDA | 5,800 | 8,700 | 12,320 |
| Current EV / Revenue | 65.4x | 49.6x | 39.1x |
| Current EV / EBITDA | 296.6x | 197.7x | 139.6x |
Even three years out (FY28E), the stock trades at ~39x forward revenue and ~140x forward
EBITDA at the current $136.12 print — a materially richer multiple than any public comp
(research.md §5: Rocket Lab ~$50.6B cap, AST SpaceMobile ~$26.2B, Planet Labs ~$9.6B, none
profitable at Starlink's segment scale). The market is pricing SpaceX as an AI-infrastructure
platform layered on a dominant, profitable launch/broadband duopoly-of-one, not as a
conventional aerospace or telecom name.
6b. Target-multiple bridge → implied 12-month price
Target EV/Revenue multiples applied to FY27E (the nearest full fiscal year, 12-18 months
forward from the 2026-07-15 as-of date) under each scenario, using each scenario's own FY27E
revenue (interpolated from FY25A to the FY30E scenario endpoints in §5) and net-cash position
held at the FY27E base-case level ($11.9B net cash) for simplicity across all three scenarios:
| Bear | Base | Bull | |
|---|---|---|---|
| FY27E revenue (scenario) | 27,413 | 34,700 | 41,640 |
| Target EV/Revenue (FY27E) | 20.0x | 35.0x | 55.0x |
| Implied EV | 548,260 | 1,214,500 | 2,290,200 |
| + Net cash (FY27E, held constant) | 11,910 | 11,910 | 11,910 |
| Implied equity value | 560,170 | 1,226,410 | 2,302,110 |
| ÷ Shares outstanding (M) | 13,000 | 13,000 | 13,000 |
| Implied price / share | $43 | $94 | $177 |
| vs. $136.12 last close (2026-07-14) | (68%) | (31%) | +30% |
6c. Read
The desk's 12-month price-target range is $43–$177/share, with a base-case point estimate
of ~$94/share — below the $136.12 last close. Even the base case assumes a compression from
today's ~50x FY27E revenue multiple to 35x, which is still a rich multiple by any conventional
standard; it is not a bearish operating forecast (base-case FY27E revenue of $34.7B is +86%
above FY25's $18.67B actual). The conclusion is a valuation call, not a fundamentals call: the
stock is currently priced closer to the bull case in this framework, leaving limited room for
multiple expansion even if the AI segment executes roughly in line with the desk's base
assumptions. This is the central input thesis.md should carry forward — the swing factor
between an Overweight and an Underweight stance is almost entirely which EV/Revenue multiple the
market is willing to sustain on the AI segment, not the core launch/Starlink business, which is
solidly profitable and growing under all three scenarios.
Sources
Primary: SEC EDGAR CIK 1181412 (S-1 2026-05-20, S-1/A 2026-06-01/03, 424B4 2026-06-12); Nasdaq
SPCX quote (checked 2026-07-15). Secondary S-1 digests: Via Satellite/satellitetoday.com
(2026-05-20), HL.co.uk, KraneShares, Vested Finance, Morningstar, X/Mo Islam compilation
(2026-05-20). Market/competitive: BryceTech via Via Satellite, SpaceNews, SatBase, Spaceflight
Now, Bloomberg, 24/7 Wall St. Full citation trail in filings.md, research.md, and the
sweep-team notes referenced throughout (financials.md, market.md, disclosure.md,
dynamics.md). All figures not carrying a specific citation above are desk estimates (E),
calibrated to tie the three statements together and to the hard S-1/Q1-2026 data points cited
inline.