How fast is satellite demand growing, who drives it, and what does the satellite catalogue show?
The short answer. Setcoin's analysis of the public satellite catalogue counts 4,492 payloads launched in 2025, against 453 in 2019 and 174 in 2015. Starlink accounted for 3,169 of them, 71%. Excluding Starlink, the rest of the world launched about 1,320 payloads in 2025, about four times its 2019 level. In 2026 to 18 October, 3,498 payloads have been launched, including 217 for Amazon Leo and 385 owned by China. Demand from everyone other than SpaceX is what meets the merchant supply chain, and it is growing at roughly 25–35% a year.
The catalogue record
Source table: Payloads launched (10 rows, shown as an exhibit above)
| Year | Payloads launched | Starlink | Amazon Leo | OneWeb | All others |
|---|---|---|---|---|---|
| 2015 | 174 | – | – | – | 174 |
| 2017 | 390 | – | – | – | 390 |
| 2019 | 453 | 120 | – | 6 | 327 |
| 2020 | 1,235 | 833 | – | 104 | 298 |
| 2021 | 1,768 | 989 | – | 284 | 495 |
| 2022 | 2,393 | 1,690 | – | 110 | 593 |
| 2023 | 2,867 | 1,984 | 4 | 132 | 747 |
| 2024 | 2,797 | 1,964 | – | 20 | 813 |
| 2025 | 4,492 | 3,169 | 182 | – | 1,141 |
| 2026 (to 18 Oct) | 3,498 | 2,194 | 217 | – | 1,087 |
Source. Setcoin space engine (CelesTrak SATCAT, 18 October 2026); payload objects by launch year; name-based grouping.
Exhibit 3.1
Output rose tenfold from 2019 to 2025; Starlink is about seven in ten
Payloads launched worldwide by year · payloads
Who owns what is launched
Source table: Who owns what is launched (7 rows, shown as an exhibit above)
| Owner (catalogue code) | 2019 | 2022 | 2025 | 2026 to 18 Oct |
|---|---|---|---|---|
| United States | 256 | 1,892 | 3,703 | 2,747 |
| China | 70 | 182 | 369 | 385 |
| United Kingdom | 7 | 111 | 7 | 8 |
| Russia | 31 | 40 | 59 | 59 |
| Japan | 10 | 3 | 22 | 32 |
| India | 9 | 11 | 10 | 5 |
| France | 5 | 5 | 25 | 8 |
Source. Setcoin SATCAT analysis. The United Kingdom's peak reflects OneWeb's deployment, registered there.
China has already launched more payloads in 2026 than in all of 2025, as its two broadband constellations, Guowang and Qianfan, ramp up (Chapter 39).
What is in orbit now
| Measure (18 October 2026) | Count |
|---|---|
| Tracked objects in orbit | 35,036 |
| Active payloads | 17,243 |
| Of which Starlink | 11,151 (64.7%) |
| Eutelsat OneWeb | 651 |
| Amazon Leo | 391 |
| Qianfan | 257 |
| Guowang | 234 |
| Active in low orbit | 16,313 |
Source. Setcoin space engine.
Replacement demand
Low-orbit satellites last five to seven years, and those in very low orbits less. A fleet of 17,000 active payloads with an average life of six years needs nearly 3,000 replacements a year before any growth. Replacement is now a larger source of demand than new constellations were in any year before 2020. This matters for supply: replacement demand is predictable, so it rewards suppliers who can commit capacity for many years, and it is concentrated in the same few constellations.
Demand the merchant chain must serve
SpaceX makes most of its satellite parts in-house (Chapter 8). Amazon makes its satellites in its own factory but buys more components from outside suppliers. Everyone else buys most parts. Setcoin therefore separates satellite demand into three streams:
Exhibit 3.2
Only the third stream buys most of its parts on the open market
Payloads launched in 2025 by supply stream
Source table: 2025 payloads (3 rows, shown as an exhibit above)
| Stream | 2025 payloads | Supply source |
|---|---|---|
| Starlink | 3,169 | Mostly in-house parts; bought semiconductors, cells, gases |
| Amazon Leo | 182 | In-house integration; mix of in-house and bought parts |
| All other satellites | 1,141 | Mostly merchant parts |
Source. Setcoin SATCAT analysis; supply characterisation from company statements and reporting.
The third stream, about 1,100–1,300 satellites a year and rising, is the demand that meets merchant suppliers of optical terminals, star trackers, wheels, thrusters and solar cells. It is the default input in the bill-of-materials model (Chapter 46).
Demand to 2030
Setcoin's demand view combines announced constellation schedules, defence tranches and replacement cycles. It does not assume every filed constellation is built.
Exhibit 3.3
Outside Starlink, demand roughly doubles to triples by 2030
Satellites a year by 2030, by demand driver (Setcoin estimate) · satellites a year
Starlink incl. replacement
EO, IoT, science, other
Chinese broadband
Amazon Leo
US defence proliferated
European sovereign
each bar runs from the low to the high estimate
| Demand driver to 2030 (Setcoin estimate) | Satellites a year by 2030 |
|---|---|
| Starlink, including direct-to-device and replacement | 3,000–5,000 |
| Amazon Leo | 600–1,200 |
| Chinese broadband (Guowang, Qianfan, others) | 500–1,500 |
| US defence proliferated architectures (SDA and successors) | 150–300 |
| European sovereign systems (IRIS² and national) | 50–150 |
| Earth observation, IoT, science and other commercial | 800–1,500 |
| Total | ~5,100–9,650 |
Setcoin estimate. Ranges reflect launch availability more than demand (Chapter 29).
Outside Starlink, demand roughly doubles to triples by 2030. That growth, concentrated in programmes that buy parts rather than make them, is the core of the supply problem this study analyses.
What would change this view
Starship flying routinely would raise Starlink demand further but would not change the merchant picture much. A collapse of one large constellation (for example, financing failure in a Chinese programme or a European budget cut) would reduce merchant demand. A conflict that destroyed satellites would create sudden replacement demand that the chain could not meet (Chapter 31).