S18.3 · Government, Defense & Public Sector
Launch, satellite manufacturing and ground infrastructure for government, moving from exquisite GEO satellites to proliferated LEO constellations.
This segment covers launch, satellite manufacturing and ground infrastructure sold to government customers. The broader global space economy is estimated at roughly $570-630B in 2023 (Space Foundation/McKinsey), projected to reach $1.8T by 2035 (WEF/McKinsey, 2024); satellite manufacturing alone runs an estimated $20-25B in 2025, growing to $114.3B by 2035 (Precedence Research, 2025) — figures that vary considerably with how broadly "space economy" is scoped. Read past the sizing noise and the structural shift is unambiguous: a small number of exquisite, cost-plus geostationary satellites is giving way to large proliferated low-earth-orbit constellations built on commercial manufacturing cadence, and that changes who earns what per unit.
Government demand centers on proliferated-constellation programs — the Space Development Agency's Transport Layer, the Golden Dome missile-defense architecture, and Phase 3 of the National Security Space Launch program — layered on commercial low-earth-orbit buildout such as Starlink and Kuiper, which is outpacing legacy geostationary demand. Buying concentrates in the United States, where the Space Force, National Reconnaissance Office and NASA write the checks, alongside New Space hubs in Los Angeles, Seattle and Northern Virginia; launch itself is physically concentrated at Cape Canaveral, Vandenberg and SpaceX's own launch sites.
Launch is a highly concentrated market — SpaceX dominant, with United Launch Alliance, Blue Origin and Rocket Lab alongside — while satellite manufacturing is more fragmented, split between legacy primes and New Space entrants. The value chain runs from launch through satellite bus and payload manufacturing to ground systems and downstream data and analytics; commercial satellite connectivity is related but separate. The economics are capital-intensive and re-pricing: cost-plus exquisite geostationary programs are yielding to fixed-price proliferated low-earth-orbit production, which moves margin risk from the government to the manufacturer and rewards production cadence over engineering exquisiteness. Proliferated-constellation demand under the Space Development Agency's tranche structure and National Security Space Launch Phase 3 now carries multi-year contracted backlog protection; satellite production contracts more broadly still live on annual appropriations. Entry is barred by ITAR munitions-list controls, classified-payload facility clearance, and FCC spectrum licensing.
The neighboring markets are downstream intelligence technology built on satellite data, the platform primes now moving into satellite manufacturing, and commercial satellite connectivity. The two camps are playing different games: legacy primes are largely protecting geostationary and exquisite-satellite franchises, while New Space entrants — SpaceX, Rocket Lab and Anduril among them — push into government space work from commercial roots. That makes this one of the sector's two clearest examples of venture-backed companies competing for prime-contractor roles, alongside homeland-security and border technology.
Agentic AI is stripping labor out of ground-segment tasking and mission planning, and autonomous constellation management is eroding the hands-on, manually operated model that protected legacy geostationary satellite operators. A distinct AI-native space-tasking and data-analytics buyer category is forming, separate from hardware manufacturing itself. The shift in launch and manufacturing economics is already underway; AI-native tasking at scale sits two to five years out.