Companies Propose Launching 8,700 Satellites into Non-Geosynchronous Orbit

OneWeb satellite. (Credit: Airbus Defence & Space)
OneWeb satellite. (Credit: Airbus Defence & Space)

While SpaceX has received most of the attention for its plan to launch more than 4,000 broadband satellite network, the constellation makes up just over half the number of spacecraft that companies have proposed placing in non-geosynchronous satellite orbit (NGSO).

Companies have filed applications with the Federal Communications Commission (FCC) to launch 8,731 NGSO communications satellites. While most of the constellations would provide broadband and communications services, others would collect Earth observation data.

According to the International Telecommunications Union, NGSO spacecraft “occupy a range of orbital positions (LEO satellites are located between 700km-1,500km from the Earth, MEO satellites are located at 10,000km from the Earth), and do not maintain a stationary position, but instead move in relation to the Earth’s surface.”

SpaceX leads the pack with 4,425 spacecraft, followed by Boeing with 2,956 and WorldVu (aka, OneWeb) with 720. Boeing has a second application before the FCC for a constellation with 60 satellites.

The table below provides a summary of the applications filed with the FCC.

SpaceXHawthorne, CA4,425Ka, KuGlobal broadband
BoeingSeattle, WA2,956VAdvanced communications, Internet-based services
WorldVu (OneWeb)Arlington, VA720KuGlobal broadband
Kepler CommunicationsToronto, ONT140KuMachine-to-machine communications (Internet of Things)
 Telesat CanadaOttawa, ONT117Ka Wide band and narrow band communications services
 Theia Holdings A, Inc.Philadelphia, PA112KaIntegrated Earth observation and communications network
Spire GlobalSan Francisco, CA100KaMaritime monitoring, meteorological monitoring, and earth imaging services
 LeoSat MAPompano Beach, FL80KaBroadband services
BoeingSeattle, WA60KaVery high speed connectivity for end-user earth stations
 O3bWashington, DC60KaBroadband services
ViaSat  Carlsbad, CA24Ka, VBroadband services
 Karousel LLCAlexandria, VA12KaCommunications
Audacy CommunicationsWalnut, CA3K, VData relay constellation providing satellite operators with seamless access to NGSO satellites
Space Norway ASOslo, Norway2Ka, KuArctic broadband





  • Jonathan A. Goff

    They didn’t file with the FCC, but Samsung has also been talking about doing a 3-4k spacecraft LEO broadband constellation. Not sure if them not filing means that they’ve dropped their concept, or are no longer in consideration. But yeah, we’re talking about a roughly order-of-magnitude increase in the number of spacecraft being proposed…


  • I’m still very much in the “it’ll never happen” camp. This feels very much like “SSTOs by the year 2000” or “suborbital tourism 2 years after the X Prize” or “pick your favorite space promise that hasn’t happened yet”.

  • Lee

    I’ll repeat a comment I made in the thread on SpaceX’s permit application:

    “So let’s do some math, shall we? Based on the dimensions of the F9
    payload fairing, and the stated dimensions here of the sats, and
    cramming the sats in, Musk might be able to launch 20 per F9. That
    comes out to 222 launches. If SX can maintain a launch rate of once per
    month for this project (remember they have other things they have to
    launch, too), we’re talking 18.5 years to get this constellation
    deployed. Assuming the same launch rate, if the fairing will only hold
    15, the number of launches goes up to 295 and time to deployment goes up
    to ~24.6 years. For 10 per fairing, the time to deployment goes up to
    ~36.9 years. If SX can’t devote one launch a month to *just this
    project*, and maintain that rate for a long time, the time to deployment
    goes up even more.

    The same arguments go for the Boeing project.

    Tell me what I’m missing that makes this project even remotely doable.”

    Note that I used the stated dimensions of the satellite body, not the sat with deployed antennas.

    Launch cadence and reliability are going to have to go a long way before either the Boeing or SpaceX constellations go from paper to reality.

  • Hug Doug

    Dimensions include deployed antennas, the sizes given are for figuring out drag for time-to-deorbit calculations. With antennas stowed, more satellites can be packed into the fairing.

  • Lee

    No, wrong. Look back at the original article on SpaceX. There is a column that shows the *body* dimensions of the spacecraft, and another that shows the dimensions of the deployed solar panels. It is not clear that “body” dimensions includes the deployed antennas. My calculations used the body dimensions.

    Even if you double the number of sats that can fit in the fairing, it’s still a ton of launches and a long time to completion of the constellation.

  • Hug Doug

    Take a good look at that chart. The title on it, specifically. “Orbital Lifetime Calculation Parameters”

    The “body” must include the deployed antenna(s), or else using that number to calculate drag over its lifetime would be useless.

    Antennas can be much larger than the body of the spacecraft, particularly for communications satellites. A safe assumption would be that deploying the antenna would at least double the size, if not triple it.

    e.g., picture.

  • Hug Doug

    To clarify the other point, solar arrays are separate because their attitude can be adjusted. They can contribute a lot to drag, or just a little.

    For example, the ISS’s Night Glider mode.

  • Lee

    Ok, cut my numbers in half. It’s still a long time until the constellation is deployed. My original argument stands.

  • Hug Doug

    It’s probably a small spacecraft bus with large deployable antenna(s). We will know for sure when SpaceX unveils their design.

    My guess is they will be made so that one launch can deploy all the satellites for a particular orbital plane.

    Per the FCC permit application, the initial deployment will consist of 32 orbital planes with 50 satellites per plane. Subsequent deployments to an additional 51 orbital planes, with 50 or 75 satellites per orbital plane.

  • WhoAmI

    Great article Doug.

  • JamesG

    What is the payload mass of a Falcon 9 v._x_ to the target orbit? That is a better denominator for your ponderings than the interior volume of their current fairing, since SX can always make a bigger one and/or Musk is not shy about trying new and crazy ideas (like a skinny “corn cob” 2nd stage).

  • JamesG

    Is obligated to file with an American regulatory agency?

  • Jonathan A. Goff

    If they’re broadcasting in the US, I think they would. I could be wrong. But OneWeb is based out of Isle of Mann, I think.


  • JamesG

    From orbit you broadcast “into” everywhere you pass over. I mean, if they were serious about building some kind of terrestrial service network, I’m sure securing a FCC license would be simple for them, and they would.

    But is a foreign company (in this case Korean) legally required (by reciprocal treaty?) to obtain a US transmitter’s permit. Just if they intend to sell the service in that nation? I don’t think other nations do…

    Honest question. Not being snarky (for a change).

  • Jonathan A. Goff

    I’m not sure either. I’m not very savvy on the telecomm regulations side of things.


  • Vultur

    I think SpaceX is assuming launch rates *much* higher than 1/month. If F9 reusability works out and they can devote manufacturing pretty much to just upper stages, and with multiple pads and a solution to helium/COPV problems, that might be possible.

    Either that or ITS will be involved…

    Boeing, I really don’t know.

  • ken gage

    This is all fantastic. SpaceX and the 4000+ satellites are most intriguing, I hope they are working on their protocol stack, the red algorithms alone will cause fits. it will certainly be a windoze free environment. Think of the reds when dealing with the various and variable delays which will be death to any app patiently waiting on the netstack to give up the packets…out of order packets will be the norm.. when they do come the app may have already moved on…. it is an environment full of interesting opportunities for genius to prevail.. dealling with the sat to sat doppler is one place rarely dealt with… cell tower to moving car doppler is covered well, but the requirement list for sats with closing speeds of thousands followed by opening speeds of thousands of fps will cause the tc in tcpip to need a bit of work.. how great for laser communication. . this topology will be awesome in space.. both control and payload would benefit from laser optical links. . I know, I’ve logged 30 or so years on both iridium and globalstar.. to include a couple patents.. hooahh… moving the cloud into a vacuum. .. the future is here. hooahh. I really hope they do get this done right before the projects are run by those who do not understand anything but political correctness and numbers on a financial spreadsheet. Engineers hope for competent engineers to both work with and for. Fear, hate, and discontent are what low caliber non-engineers bring to the show.
    the debris will be less than one would think, 4000 more satelites will be more work getting them up than managing them in mission orbit.. Ever thought about the elbow room space has and how it exponentially increases with an ever lengthening radius? staggering. .. management increases with more stuff,
    but more targets for various asat programs will be a factor as well.
    lots and lots of room in space for big brother to be flummoxed and be over taken by lots of targets or rather commercial endeavors.
    wow… who pulled my chain…?
    peace out. hooahh

  • Colin Helms

    Then there was the Commercial Mission Model back in 92. Iridium went belly up and caused the “Nuclear Winter of Satellites”. Let’s just put up a big hose and fill near space with garbage; it comes to the same thing, but will make business for someone with an idea about orbital debris removal. It’s ridiculous to talk about nits like protocol stacks. All been there, done that in the real space comm industry. Only three things really count; utility, visibility, and lifetime. Getting spectrum for comm is a bitch too. It’s easy to put up a nose picking cube sat run powered by a transistor radio battery. Lifetime of 3 months if it doesn’t collide with it’s sister within the first 7 orbits.

  • Euripedes_Smythe

    Agree with a point that Colin Helms makes below — remember Iridium, remember Iridium, remember Iridium!!!