of the satellites the world’s flying operators have announced are in orbit today.
How far the global constellation buildout has come — counting only operators that already have a satellite flying and a fleet target on file with their regulator (FCC, ITU or equivalent). Plenty more have announced constellations but aren’t counted yet, until they put a satellite up or get their target approved — so the buildout keeps growing as they qualify.
Bars are real, reconstructed from GCAT launch/decay dates; the target is today’s figure (no back-history), shown flat. Buildout counts one-time deployment, not replenishment. Full detail in How I count this below.
Most of the remaining buildout is flown in-house or by the state and never priced on the open market — so a single “cost to build out” number would mislead. The honest cut is the launch demand that’s genuinely bought.
Starlink → SpaceX; Guowang / Qianfan → China’s state rockets. Flown at internal / subsidised cost — not an open-market transaction, so not priced here.
3,083 sats bought on the commercial market — Kuiper, AST SpaceMobile and others — priced at list. This is the addressable market; real contracts come in lower.
7 more constellations have an announced target but aren’t counted yet — they join the buildout once they launch a satellite or get their target approved. Filed systems still on the ground are on the Watchlist (E-Space, Rivada, Mangata); pre-launch constellations with approved plans (Telesat Lightspeed, IRIS², Open Cosmos) sit in Constellations.
How far the world’s flying satellite operators are toward their announced fleets — planned vs in orbit — plus the launches, years and third-party launch-market cost still ahead. Buildout only: the one-time launches to reach the targets, not the recurring launches that replace satellites at end of life.
Buildout = active ÷ planned, summed across the constellations whose operator is actually flying — status operational, mature or building AND at least one satellite in orbit. That gate excludes paper filings (E-Space, Reflect) and pre-launch constellations (Telesat Lightspeed) until they start deploying. Planned fleet sizes are hand-curated with a source per figure; active counts come from CelesTrak.
Planned = the operator’s announced / licensed operational target, NOT its maximal speculative ITU spectrum filing. For US operators that’s the FCC-authorised system — Starlink 15,000 (its Gen-2 licence as of Jan 2026, not the ~42,000 it has filed at the ITU); Amazon Leo 3,236 (the Gen-1 licence it’s deploying). For others it’s the declared full-system size — Guowang 12,992 as filed with the ITU, which is China SatNet’s stated system with no separate lower licence. Using the licensed figure keeps the buildout gap tied to what’s actually being built, not to spectrum placeholders.
Remaining satellites (planned − active, floored at zero) ÷ each constellation’s dedicated-launch batch size, summed across the biggest dedicated-launch constellations only. Batch sizes are hand-curated from demonstrated flights with a source, NOT read from the GCAT join — the join reports ~1 satellite per launch for Guowang (messy naming) and ~3 for rideshare operators, which would invent tens of thousands of phantom launches. Rideshare and small operators, which don’t buy dedicated launches, are excluded from the launch count entirely.
For the dedicated-launch constellations that BUY launches on the open commercial market, launches-needed × the list price of the rocket they use (launcher catalog). Excludes vertically-integrated launches (SpaceX flies its own Starlink) and state launches (China’s national megaconstellations on state rockets), which aren’t market transactions. Shown at list prices as an order-of-magnitude figure — real contracts are negotiated below list. The single biggest uncertainty is fairing-limited batch sizes (AST SpaceMobile).
Remaining ÷ the most recent full calendar year’s net fleet growth (year-end active minus the prior year-end, across the flying operators). A rough runway, not a forecast — it assumes today’s deployment rate holds.
These count only launches that GROW a fleet to its announced target — a finite, one-time buildout. They exclude replenishment: the recurring launches that REPLACE satellites at end of life (~5 years in low orbit). A mature constellation like OneWeb reads ‘complete’ here even though it keeps launching to stay alive; Eutelsat’s 440 next-gen OneWeb satellites are replenishment (they keep the fleet at ~648, not grow it) and don’t count. Replenishment is a separate, larger, recurring market we may track later.
Satellites in orbit at each year-end for the flying operators, reconstructed from GCAT launch and decay dates (real back-history, available now). The announced target is today’s figure, shown flat — planned sizes are curated as-of-now with no history. The current year is a partial (YTD) bar. A finer captured-historical series accrues forward from daily snapshots (data_snapshots) and layers in over the coming months.
Everything is computed at today’s capabilities — current rockets, batch sizes and list prices; heavier lift (Starship, New Glenn) would compress launches, years and cost. Announced fleet sizes are operator aspirations, not commitments, and many will change. The totals tend to RISE over time as more constellations reach operational status and operators announce new targets, even as today’s are filled in — so a step up isn’t necessarily real-world growth. Batch sizes and the third-party market classification are curated judgment calls, documented in the code.
Computed by Open Orbit. Underlying data: CelesTrak (attribution) + GCAT (CC-BY).