SPCX is in our analyst coverage. Read the full analyst research →

Signal Ledger valuation · SPCX

Space Exploration Technologies Corp

What the shares are worth today on our research (cited consensus, company plans and peers, since management sets no usable targets, plus our judgment calls): every line valued at the end of FY2030, net debt rolled forward, then discounted back to Sep 30, 2026.

Signal Ledger valuation$139.04−4% vs $144.75Includes 3 judgment calls of ours · see them ↓
Business value at FY2030
$3,469B
FY2030 revenue
$652B
FY2030 EPS · P/E
$8.69 · 24×
Discount rate (WACC · equity)
9.26% · 9.00%

How this is built. Management sets no usable targets, so each business is built from cited consensus, company plans and peers, with the judgment calls listed below. Each input below shows its basis: consensus, company plan, peer or rule.

Researched Oct 23, 2026 from SEC filing 0001628280-26-066880 · price $144.75 at research time · Valuation methodology

Latest quarter in this research: Q3 2026

Reported Oct 20, 2026 for the quarter to Sep 30, 2026 (company release ↗). This valuation already reflects these results (researched Oct 23, 2026).

Revenue$14.6B
Operating income$2.79B19.1% margin
Net income$4.40B
Diluted EPS$0.35
SegmentRevenuevs a year agoOperating incomeMargin
Super Intelligence (AI)$7.56B+830%$0.43B5.7%
Connectivity (Starlink)$5.25B+69%$2.37B45.1%
Space (launch)$1.77B+6%−$7M-0.4%
  • Super Intelligence revenue tripled from Q2 to $7.56B on new cloud services agreements, Cursor (closed August 14) and Grok, and turned an operating profit of $0.43B after $1.2B of Cursor stock compensation.
  • Starlink reached 13.6 million consumer subscribers (+1.6 million in the quarter) at $69 a month; Connectivity earned a 45% operating margin.
  • Nameplate compute reached 1.7 GW (1.4 GW in June); capex was $19.6B in the quarter, $16.4B of it for AI.
  • Starship flew its first revenue mission, carrying 26 Starlink V3 satellites.

What each business is worth

BusinessModelFY2030 revenueValue todayPer share2030 multipleValue from beyond FY2030
Starlink and launchCapacity (capex follows demand)$152B$794B$58.4712.7× EBITDA79%
Super Intelligence (AI)Capacity (capex follows demand)$500B$1,042B$76.708.4× EBITDA109%
Digital assets (bitcoin)treasury digital assets at fair valueMarket mark—$1.6B$0.12——
Businesses and holdings$652B$1,837B$135.29

Starlink and launch: does new capacity pay?

Each $1 of equipment earns 66.7% EBITDA a year against a 24.0% break-even. Every $1 of new annual revenue is worth $1.35 today after paying for its equipment.

Needs $42B of funding before it pays for itself; that lands in net debt.

Super Intelligence (AI): does new capacity pay?

Each $1 of equipment earns 28.9% EBITDA a year against a 21.6% break-even. Every $1 of new annual revenue is worth $0.55 today after paying for its equipment.

Needs $721B of funding before it pays for itself; that lands in net debt.

Our judgment calls

Where the evidence leaves room for judgment, we make the call ourselves and say why (Signal Ledger, Oct 20, 2026). These 3 calls are part of the valuation above; every other input follows the research file and the published rules.

AssumptionOur callWhy
Discount rate (cost of equity)9.0%SpaceX is treated as market-average risk (beta about 1): 3.85% normalised risk-free rate + 1.0 × 5% equity premium ≈ 9%, the same call as for AMD, instead of the 10% the beta cap alone would give; WACC, which carries a build-out premium, moves down one point with it.
Starlink and launch: FY2030 revenue$152BGoldman Sachs' 2030 forecast (Starlink $144B + rockets $8.3B, reported by Yahoo Finance: https://finance.yahoo.com/markets/stocks/articles/goldman-sachs-predicts-spacex-revenue-163126173.html). Q3 2026 already runs at $28B a year with Connectivity +69% and enterprise and government +91%; Starlink Mobile on the new 800 MHz spectrum adds a market beyond home and enterprise broadband.
Super Intelligence (AI): FY2030 revenue$500BSuper Intelligence grew 195% in Q3 2026 to a $30B run-rate and turned profitable at the operating line (8-K ex 99.1, 2026-10-20). $500B in 2030 is ~28 GW of compute at Q3's ~$17.8B of revenue per GW-year, above Goldman's $322B: it assumes the orbital-compute and Colossus build-outs keep the 2026 pace and that Grok and Cursor grow software revenue on top of leased capacity.

From 2030 back to today

  1. Enterprise value at end of FY2030$3,469B
  2. Less net debt at FY2030$656B
  3. Equity value at FY2030$2,814B
  4. ÷ (1 + 9.00%)^4.3 years at the cost of equity$1,951B
  5. + dividends paid on the way, discounted$0B
  6. + investments held outside the businesses$1.6B

This reconciliation checks the published Signal Ledger valuation against the net-debt path. The difference is 3.4%, inside our 5% tolerance.

From businesses to one share

  1. All businesses, valued today$1,836B
  2. − borrowings$38B
  3. − finance leases$1B
  4. + cash and short-term investments$90B
  5. + investments held outside the businesses$1.6B
  6. ÷ diluted shares13.581B
  7. Signal Ledger valuation$139.04

Year by year to FY2030

Fiscal yearRevenueEBITDACapexFree cash flowNet debt at year end
FY2026$44B$14B$68B−$55B$38B cash
FY2027$83B$31B$133B−$105B$64B
FY2028$160B$70B$272B−$208B$277B
FY2029$320B$160B$579B−$437B$731B
FY2030$652B$366B$201B$122B$656B

What must be true

  • Starlink and launch grows revenue from $23.2B in FY2026 to $152.3B in FY2030 (6.6×).
  • Super Intelligence (AI) grows revenue from $20.4B in FY2026 to $500.0B in FY2030 (24.5×).
  • Starlink and launch raises its EBITDA margin from 33.5% to 60.0%.
  • Super Intelligence (AI) raises its EBITDA margin from 29.5% to 55.0%.
  • The company spends $1252.7B of capital expenditure through FY2030.
  • Net debt peaks at $731.0B in FY2029 and stays financeable.
  • Starlink and launch earns 66.7% EBITDA a year on each dollar of equipment, above the 24.0% it needs to cover depreciation and the cost of capital.
  • Super Intelligence (AI) earns 28.9% EBITDA a year on each dollar of equipment, above the 21.6% it needs to cover depreciation and the cost of capital.

What the price implies

To reach $144.75 at our discount rate, Space Exploration Technologies Corp would need about $681B of FY2030 sales (growth of about 98.8% a year from FY2026) with 56.2% EBITDA margin — versus $652.3B and 56.2% EBITDA margin in our research.

At the research price of $144.75 (2026-10-23), the price asks more than our assumptions overall. These are operating requirements at that price, not additional valuations or forecasts.

If sales and margins changed together

Each row matches the same research price at our unchanged discount rate, by scaling every operating line's margin by one common factor. Sales cover margin and capacity lines only; other valuation lines stay fixed. Margins are revenue-weighted within each metric, before separately modeled corporate costs and eliminations. The conclusion uses the row at our research margins when sales alone can match the price, otherwise the solved row needing the least sales growth; this is a trade-off between growth and profitability, not a forecast.

FY2030 sales factorFY2030 salesFY2026–30 CAGRRequired FY2030 marginRequirement
1.04×At research margins$681B98.8%
56.2% EBITDA
Within margin guardrails
1.50×$978.4B117.7%
49.9% EBITDA
Within margin guardrails
2.00×$1,304.6B133.9%
46.3% EBITDA
Within margin guardrails
3.00×$1,956.9B158.9%
42.7% EBITDA
Within margin guardrails

CAGR uses $43.6B of same-scope FY2026 sales from the published model or its sourced research base, over four fiscal years. “Within margin guardrails” checks numeric limits only, not commercial plausibility. Each line and year must stay at or below 50.0% GAAP EBIT and 100% EBITDA, or our research margin where that is higher; capacity EBIT uses the engine's equipment depreciation. Unsolved margins are not extrapolated.

Implied discount rate
9.1%

The WACC that makes model value equal the price, shifting WACC and cost of equity by the same amount. Model: 9.3% WACC · 9.0% cost of equity.

All-business FY2030 sales factor
1.04×

FY2030 revenue would need to reach $681B (model: $652.3B) across the margin and capacity businesses. FY1 stays fixed; the sales change ramps to FY2030 and stays at that factor thereafter.

If only one business changed

Businesses with margin or capacity revenue paths, ordered by value today. Revenue and margin are separate alternatives; these requirements are not added together.

Super Intelligence (AI)

Revenue alone
FY2030 revenue would need to reach $539.9B (model: $500B).
Margin alone

FY2030 EBITDA margin would need to reach 56.2% (model: 55.0%).

Starlink and launch

Revenue alone
FY2030 revenue would need to reach $167.4B (model: $152.3B).
Margin alone

FY2030 EBITDA margin would need to reach 63.8% (model: 60.0%).

Mechanical requirements, not forecasts. The separate sales-only and single-business factors are searched from 0.1× to 6×; an unsolved comparison means this business alone cannot get there within those bounds. Single-business revenue changes apply from FY2 onward; margin changes scale the full EBIT or EBITDA margin path. Mixed groups show revenue-weighted EBIT and EBITDA margins separately. Requirements above 100% are not feasible margins. Joint rows instead search non-negative uniform margin factors only up to the EBIT/EBITDA guardrails; the entire margin path scales, including post-FY2030 convergence. Sales and margins change together only in that table. The rate search runs from the largest terminal growth rate plus 0.5 percentage points to 30% WACC.

The discount rate

Cost of equity
9.00%

Signal Ledger judgment: SpaceX is treated as market-average risk (beta about 1): 3.85% normalised risk-free rate + 1.0 × 5% equity premium ≈ 9%, the same call as for AMD, instead of the 10% the beta cap alone would give; WACC, which carries a build-out premium, moves down one point with it.

WACC
9.26%

Signal Ledger judgment: SpaceX is treated as market-average risk (beta about 1): 3.85% normalised risk-free rate + 1.0 × 5% equity premium ≈ 9%, the same call as for AMD, instead of the 10% the beta cap alone would give; WACC, which carries a build-out premium, moves down one point with it.

Every input, with its source

Show 23 inputs · reported, guidance and rule-based
InputValueBasisWhy
Revenue FY2026Starlink and launch$23.2BReportedtarget (methodology §4.2), growth 49.6% sec.gov ↗ · finance.yahoo.com ↗
Revenue FY2030Starlink and launch$152.3BRuleSignal Ledger judgment: Goldman Sachs' 2030 forecast (Starlink $144B + rockets $8.3B, reported by Yahoo Finance: https://finance.yahoo.com/markets/stocks/articles/goldman-sachs-predicts-spacex-revenue-163126173.html). Q3 2026 already runs at $28B a year with Connectivity +69% and enterprise and government +91%; Starlink Mobile on the new 800 MHz spectrum adds a market beyond home and enterprise broadband. Our judgment calls · Read methodology ↗
EBITDA margin FY2030Starlink and launch60.0%Consensuslinear path to 60.0% by FY2030 (management target) — operating margin 45.0% + depreciation k ÷ life = 0.9 ÷ 6 = 15.0% of revenue → EBITDA 60.0% finance.yahoo.com ↗ · sec.gov ↗
Equipment per $1 of revenue (k)Starlink and launch0.9Rulemethodology §4.4: in-service equipment + FY1 capex over FY2 revenue = 0.57, averaged with the contract 1.2 sec.gov ↗
Equipment lifeStarlink and launch6 yearsRuleserver life 5 years + 1 for long-lived assets SpaceX: Starlink satellites are designed for about five years in orbit (starlink.com/technology)
Capex FY2026Starlink and launch$11.4BReported2026 company capex: nine months plus a fourth quarter at Q3's level × 0.168 pool share sec.gov ↗
Terminal growthStarlink and launch3.0%Rulemethodology §4.2 terminal growth for a growth line Valuation methodology §4.2 · Read methodology ↗
Revenue FY2026Super Intelligence (AI)$20.4BReportedtarget (methodology §4.2), growth 538.0% finance.yahoo.com ↗ · fortune.com ↗ · sec.gov ↗
Revenue FY2030Super Intelligence (AI)$500.0BRuleSignal Ledger judgment: Super Intelligence grew 195% in Q3 2026 to a $30B run-rate and turned profitable at the operating line (8-K ex 99.1, 2026-10-20). $500B in 2030 is ~28 GW of compute at Q3's ~$17.8B of revenue per GW-year, above Goldman's $322B: it assumes the orbital-compute and Colossus build-outs keep the 2026 pace and that Grok and Cursor grow software revenue on top of leased capacity. Our judgment calls · Read methodology ↗
EBITDA margin FY2030Super Intelligence (AI)55.0%Peerlinear path to 55.0% by FY2030 (management target) sec.gov ↗
Equipment per $1 of revenue (k)Super Intelligence (AI)1.9Rulemethodology §4.4: in-service equipment + FY1 capex over FY2 revenue = 1.65, averaged with the contract 2.1 sec.gov ↗
Equipment lifeSuper Intelligence (AI)7 yearsPeerserver life 6 years + 1 for long-lived assets CoreWeave and Microsoft depreciate GPU servers over six years (10-K useful-life policies)
Capex FY2026Super Intelligence (AI)$56.6BReported2026 company capex: nine months plus a fourth quarter at Q3's level × 0.832 pool share sec.gov ↗
Terminal growthSuper Intelligence (AI)3.0%Rulemethodology §4.2 terminal growth for a growth line Valuation methodology §4.2 · Read methodology ↗
WACCDiscount rates9.26%RuleSignal Ledger judgment: SpaceX is treated as market-average risk (beta about 1): 3.85% normalised risk-free rate + 1.0 × 5% equity premium ≈ 9%, the same call as for AMD, instead of the 10% the beta cap alone would give; WACC, which carries a build-out premium, moves down one point with it. Our judgment calls · Read methodology ↗
Cost of equityDiscount rates9.00%RuleSignal Ledger judgment: SpaceX is treated as market-average risk (beta about 1): 3.85% normalised risk-free rate + 1.0 × 5% equity premium ≈ 9%, the same call as for AMD, instead of the 10% the beta cap alone would give; WACC, which carries a build-out premium, moves down one point with it. Our judgment calls · Read methodology ↗
Tax rateDiscount rates21.0%Rulestatutory rate; net operating losses not modelled 26 U.S.C. §11(b)
Cash and short-term investmentsBalance sheet$89.9BReportedcash, cash equivalents and short-term investments us-gaap:CashAndCashEquivalentsAtCarryingValue + us-gaap:MarketableSecuritiesCurrent · 0001628280-26-066880 (filed 2026-10-22)
BorrowingsBalance sheet$37.9BReportedborrowings due within a year plus long-term borrowings; finance leases already included in combined debt tags removed and deducted separately us-gaap:LongTermDebtAndCapitalLeaseObligationsCurrent · 0001628280-26-066880 + us-gaap:LongTermDebtAndCapitalLeaseObligations · 0001628280-26-066880
Finance leasesBalance sheet$1.0BReportedfinance lease liabilities (capex-like; deducted once) us-gaap:FinanceLeaseLiability · 0001628280-26-066880 (filed 2026-10-22)
SharesBalance sheet13.581BReportedshares outstanding as of 2026-10-15 cover page of 10-Q 0001628280-26-066880 (2 share classes summed)
Digital assets (bitcoin)Non-operating assets$1.6BReportedtreasury digital assets at fair value sec.gov ↗
Dividends per yearShareholders$0.0BReportednone × 13.581B shares no dividends

Research notes

  • share-count comparison: 12.475B quarter 2026-07-01..2026-09-30 · 0001628280-26-066880; point-in-time denominator retained
  • CAPEX_HISTORY assumptions (USD billions): 2019: 4.415000, 2020: 4.415000, 2021: 4.415000, 2022: 4.415000. FY 2019 reconstructed from earliest FY 2023: capex $4.415B / revenue $10.387B (42.5051%); no vintage revenue: hold anchor capex constant; public history shorter than asset life; SEC 0001628280-26-042639; Valuation methodology §9. Each vintage retains its own basis / ref / reason in the model.
  • cost of equity 9.87% (β 1.20 after Blume adjustment), WACC 10.1% incl. 0.332248-point build-out premium (§4.5 G2)
  • Starlink and launch: k 0.9 (calibrated 0.57, contract 1.2), life 6 years
  • Starlink and launch: operating-margin target 45.0% converted to 60.0% EBITDA (k ÷ life 15.0%)
  • Super Intelligence (AI): k 1.9 (calibrated 1.65, contract 2.1), life 7 years
  • §4.5 G2: FY1 net debt increase $13.063041780000006B / (2% × equity market value $1965.85897187775B); build-out premium 0.332248pp, measured before premium

Signal Ledger research built from cited consensus, company plans and peer economics, with our stated judgment calls, because management sets no usable targets — not a forecast and not investment advice. The research desk (Claude Opus 5.5) gathers filings, guidance and targets; deterministic code applies the samepublished rules to every company and checks the result. Inputs may contain errors; follow the source links before relying on them. Read our valuation methodology →