Industry
Industry — Understanding the Arena
1. Industry in One Page
SpaceX is filed under Aerospace & Defense (GICS Industrials), but the label hides the real arena. The company operates in the modern space economy — a $626B global market in 2025 (Novaspace), projected to reach $1T by 2034 — which fuses three previously separate industries: orbital launch, satellite communications, and (since the early-2026 xAI acquisition) frontier AI compute. The unifying economics are simple: lower the cost of putting mass into orbit and you create the option to build new industries on top of that mass.
Legacy Aerospace & Defense is cost-plus, low-margin, cycle-driven manufacturing (Lockheed Martin's FY2025 gross margin was 10.2%). The new space stack is infrastructure-platform economics (Starlink's connectivity segment ran a 38.8% operating margin in FY2025). Investors who anchor on defense-prime multiples will miss the shape of the cash flows here.
Mental model: Read SpaceX as a launch monopoly that has used its cost advantage to forward-integrate down the stack — first into a global broadband ISP (Starlink), then into AI compute (xAI). The launch layer is the moat; the layers above it are where the cash compounds.
2. How This Industry Makes Money
The economics differ sharply by layer. Launch is a heavy-capex, low-volume hardware business where the unit is the rocket flight and price is quoted per kilogram to orbit. Satellite manufacturing is a fixed-cost factory business where unit economics improve with volume. Connectivity services is a subscription business where ARPU times subscribers minus network depreciation drives margin. Defense contracting is cost-plus or firm-fixed-price work, often with restrictive data rights but low offtake risk.
The key insight: launch is loss-leader infrastructure for the layers above. SpaceX's S-1 makes this explicit — the Space segment is run "to support our businesses and those of our customers," meaning internal Starlink and (soon) AI compute launches generate no segment revenue but consume the bulk of cadence. Only third-party customer launches count as Space revenue. That is why a 2,213-metric-ton mass-to-orbit year in 2025 produced just $4.1B in reported Space revenue while delivering most of the value internally.
Connectivity is the cash engine; Space is the moat that feeds Connectivity; AI is the loss-making bet that consumes 60% of the FY2025 capex pie. Expect this profit-pool shape — concentrated in one segment, with adjacent segments running at investment-grade losses — to persist for several years.
3. Demand, Supply, and the Cycle
This is not a single-cycle industry. Each layer cycles on different drivers, and they only weakly correlate.
A downturn here looks unusual. Pricing pressure shows up first in subscriber ARPU (Starlink's ARPU fell from $86/mo in Q1-25 to $66/mo in Q1-26 as international mix grew). Volume softness shows first in customer launches (down to 7 in Q1-26 from 12 in Q1-25, even as overall cadence rose). Defense budget shocks show first in program-of-record decisions, not topline.
Supply-side fragility worth flagging. Reporting in early 2026 documents tight supply of space-hardened electronics, focal planes, AI processors, helium and xenon. SpaceX's 10,000+ supplier network gives it scale leverage but also exposure to single-source bottlenecks that the SpaceX vertical-integration thesis is partly designed to escape (e.g. Raptor engines built in-house, Starlink terminals built in Bastrop, Texas, Terafab chip JV with Tesla/Intel).
4. Competitive Structure
The arena is concentrated at the launch layer, fragmented at the connectivity layer, and oligopolistic at the defense layer. SpaceX is the rare company that competes in all three at once. Public comparables exist only at the edges.
SpaceX captured more than 80% of the world's mass to orbit in 2025 and conducted ~80% of U.S. launches. By cadence (165 Falcon launches in 2025 vs. ~10 from ULA and ~18 from Rocket Lab), the launch industry now has one provider operating at airline frequency and a peloton of providers measured in single-digit monthly attempts.
The connectivity layer is the opposite. Starlink's ~10.3M subscribers compete with terrestrial telcos with hundreds of millions of customers (Verizon, Comcast, T-Mobile, Vodafone), GEO satcom incumbents (EchoStar/Hughes, Viasat, SES), other LEO entrants (Amazon Kuiper, Eutelsat OneWeb, Iridium), and direct-to-cell challengers (ASTS, Lynk, Globalstar). No single competitor is a clean comparable.
SpaceX's FY2025 revenue is larger than its four listed peers combined ex-Lockheed. Its valuation (private-market reference: $1.51T per Forge as of early May 2026; reported IPO target $1.75T per Reuters) sits closer to a hyperscaler than to any A&D peer. The peer set is a mosaic of segment proxies, not a clean comp.
Any "competitive intensity" view that ignores Blue Origin, ULA, and Project Kuiper materially understates the threat set, because none of those firms file US 10-Ks.
5. Regulation, Technology, and Rules of the Game
This industry runs on government licenses. Three regulators dominate, and a fourth — international spectrum coordination via the ITU — can stall a constellation for years.
The two regulatory levers most likely to move estimates are FAA Starship licensing throughput (gates AI compute capex, since orbital data centers require Starship cadence) and FCC spectrum coordination (gates Starlink Mobile growth, which is the highest-incremental-margin growth line). Both are visible in real time through filings on the FAA's commercial space dashboard and the FCC's IBFS database.
6. The Metrics Professionals Watch
Equity ratios alone do not explain this business. The right scorecard mixes operational, regulatory, and unit metrics.
The compound metric a senior analyst would track first is mass-to-orbit per dollar of capex deployed (a rough proxy for launch productivity) crossed with subscriber net adds per gigabit of constellation capacity (network monetization). Neither is reported directly; both can be derived from the S-1.
7. Where SpaceX Fits
SpaceX is not an A&D incumbent and is not a pure-play launch challenger. It is a vertically integrated platform that occupies multiple positions on the value chain at the same time.
The implication for the rest of the report: SpaceX cannot be valued on a single peer multiple. The Connectivity segment ($11.4B FY25 revenue at 38.8% operating margin) behaves like a fast-growth telecom/SaaS hybrid. The Space segment is the moat-creating infrastructure that justifies internal cross-subsidies. The AI segment is an early-stage hyperscaler-style bet that consumed 60% of FY25 capex for 17% of FY25 revenue. Warren's Business tab and the valuation pages will need to model these separately rather than apply an EV/sales blender.
8. What to Watch First
A reader who only has fifteen minutes a month for this name should track this short list. Each signal is observable in filings, regulatory dashboards, or credible industry press.
The single most informative cross-check. Watch the ratio of internal Starlink launches to customer launches. In 2025 it was roughly 122 internal to 43 customer (2.8x). When this ratio rises sharply, SpaceX is shifting capacity to build out its own constellation faster than serving third parties — bullish for Connectivity revenue 12 months out, bearish for Space-segment optical revenue today. When it falls, the launch market is tightening and external prices follow.