More than 37,000 satellites are forecast to launch between 2023 and 2033, according to Analysys Mason. That is a factory problem as much as a launch problem.
The factories that feed them are the least glamorous part of the orbital economy. They are now the most binding constraint. Five years ago the bottleneck was launch. Today it is how fast spacecraft can be built.
That question is splitting the industry into two answers.
The industry is splitting into two models: deep in-house integration (SpaceX, Apex pushing toward full vertical ownership) and open modular supply (EnduroSat, Swissto12, Kongsberg NanoAvionics).
The survivors in the open market are niche specialists, not general-purpose integrators: mission-specific spacecraft, small geostationary (GEO) satellites, and component suppliers.
In May, MDA Space opened a 185,000-square-foot satellite factory in Montreal built to handle up to 400 spacecraft a year, two years after breaking ground. In June, Apex raised $200 million at a $2.3 billion valuation to expand its Los Angeles plant. In August, Muon Space took $250 million in a round that valued it at $1.5 billion, with a new San Jose facility aimed at 500 satellites a year by 2027.
The capital is not going to clever payloads. It is going to floorspace, cleanrooms, and assembly tooling. Analysys Mason puts cumulative satellite manufacturing and launch revenue at $723 billion between 2023 and 2033, and the spending is front-loaded into the next three years.
The integrated factory
Vertical integration has become the default answer to the throughput question. SpaceX built Starlink in-house because no supplier existed at the scale it needed. The logic has now spread to the newest entrants.
Apex's Aries bus, the smallest of its three platforms, is about 30% built in-house. Its mid-range Nova platform is more than 70% in-house, and the company says it will move to "pretty much 100% vertical integration as we continue to scale up."
"I think we're seeing in the industry, the successful players are increasingly vertically integrated," said Armand Musey, founder of the satellite industry analyst Summit Ridge Group.— Armand Musey, Summit Ridge Group (SpaceNews)
In-house manufacturing buys three things: speed, volume, and control over the supply chain. When a program needs hundreds of identical buses on a fixed schedule, owning the line removes a negotiation step between every subsystem. Muon's San Jose factory and MDA's Montreal plant are the same bet at larger scale.
There is a cost, and it is capital. A factory built for 400 satellites a year needs an order book to justify itself. Integrated players are making a fixed-cost commitment to a demand forecast. If the forecast wobbles, the factory still has to be paid for.
The open supplier market
The counter-model treats satellite production as a supply chain problem, not a factory problem. Standardize the bus, let operators configure payloads, and buy subsystems from whoever is best at making them.
Swissto12 raised $70 million in July to accelerate production of its small GEO satellites, holding more than $500 million in contracts. The Swiss firm reports 110% compound annual growth since 2022 and keeps engineering, manufacturing, and test facilities under one roof for fast iteration. Kongsberg NanoAvionics and EnduroSat run the same playbook at smaller mass classes.
"Most vertical integrations are the result of being hurt by a supplier," said Viktor Danchev, chief technology officer of EnduroSat. "If you're very good at doing it yourself, you can consider vertical integration."— Viktor Danchev, EnduroSat (SpaceNews)
The supplier market matured because defense procurement programs, led by the Space Development Agency's proliferated architecture, invested in a base of independent bus makers. Analysys Mason notes the beneficiaries are the component suppliers: a component maker "can be more agnostic, where you're not as locked in," the way an integrator is.
The opening in this model is the one-off. Large manufacturers deprioritize mission-specific spacecraft, and standardized platforms cannot serve every need. Abhishek Tripathi, director of mission operations at the University of California's Space Sciences Lab, calls it "a tremendous opportunity to be a high-quality provider of mission-specific satellites."
The data behind both arguments
Satellite launch pipeline
Forecast number of satellites launched between 2023 and 2033, most of them for communications constellations. · Analysys Mason, 2024
Satellite manufacturing + launch revenue
Cumulative forecast for building and launching satellites, with spending front-loaded into the next three years. · Analysys Mason, 2024
Vertical integration at Apex
Share of the mid-range Nova bus developed in-house, heading toward full integration as the company scales. · SpaceNews, 2026
| Parameter | In-house factory | Open supplier market |
|---|---|---|
| Speed at scale | ✔ hundreds of identical buses | ◐ small series, configurable |
| Up-front capital | ✗ factories, cleanrooms, tooling | ✔ uses existing capacity |
| Supply-chain control | ✔ full ownership | ✗ depends on vendors |
| Unique missions | ✗ deprioritized | ✔ the core niche |
| Best fit | mega-constellations, defense programs | regional operators, sovereign programs |
Two production models compared. Sources: SpaceNews, Analysys Mason, company disclosures, 2026.
Neither model is static. MDA is already automating assembly with augmented reality and proprietary test chambers to cut validation time. Swissto12 is adding floorspace and pulling computer numerical control (CNC) machining in-house even while selling on the open market. The practical answer most players are converging on is selective integration: own the step that gates your schedule, buy everything else.
✔ Arguments for the integrated factory
+ Volume: hundreds of satellites a year per factory
+ Control: no negotiation step between subsystems
+ Value capture: the margin stays inside the company
Confirmation criteria: integrated factories hit production targets ahead of schedule while keeping unit costs flat.
✗ Arguments against
− Demand-forecast dependency: capacity is a bet on a projection
− Supplier ecosystem erosion: fewer independent vendors over time
− Mission flexibility: one-offs get deprioritized
Disconfirmation criteria: a vertically integrated ramp slips while a supplier-driven competitor ships on time.
The two models are not symmetrical. Integration is winning the headline rounds: the funding, the factories, the large defense orders. But the open market is not disappearing. It is retreating to the edges where integrated players cannot follow, serving mission-specific spacecraft, regional operators, and sovereign programs that will not hand their supply chain to a single company.
As we wrote in September, K2 Space raised $500 million to scale its satellite line. The factories keep getting bigger. The question is who still owns a line when the constellation orders slow down.