Future of SpaceX Starlink and Starships

SpaceX has submitted requests to add 30,000 satellites for Starlink 3.0. They had 20 filings to the ITU (International Telecommunications for 1,500 satellites apiece in various low Earth orbits. SpaceX had originally filed for an initial Starlink network of 4425 satellites. Once the first 1000 satellites are up they will be able to start operation.

Starlink Constellations

Starlink 1.0   4,425 satellites around the end of 2021
Starlink 2.0   7,518 satellites around the end of 2023
Starlink 3.0  30,000 satellites around the end of 2027

In November 2018, SpaceX received US regulatory approval to deploy 7,518 broadband satellites, in addition to the 4,425 approved earlier. SpaceX’s initial 4,425 satellites had been requested in the 2016 regulatory filings to orbit at altitudes of 1,110-kilometer (690 mi) to 1,325-kilometer (823 mi). The new approval was for the addition of a very-low Earth orbit NGSO [non-geostationary satellite orbit] constellation, consisting of 7,518 satellites operating at altitudes from 335-kilometer (208 mi) to 346-kilometer (215 mi). In November, SpaceX altered the orbits of 1,600 from 1,150 km (710 mi) to only 550 km (340 mi) orbital altitude.

In April 2019, the FCC approved place nearly 12,000 satellites in three orbital shells: initially approximately 1,600 in a 550-kilometer (340 mi)-altitude shell, and subsequently placing ~2800 Ku- and Ka-band spectrum satellites at 1,150 km (710 mi) and ~7500 V-band satellites at 340 km (210 mi).

The next 30,000 satellites would operate in low Earth orbit at altitudes ranging from 328 kilometers to 580 kilometers.

Elon Musk would leverage the Tesla experience of scaling to the production of 500,000 electric cars to scale up satellite mass production.

SpaceX Launch Timing

SpaceX launched 60 Starlink 0.9 prototypes in a Falcon 9.

The 60 Starlink v0.9 satellites, launched May 2019, have the following characteristics:

Flat-panel design with multiple high-throughput antennas and a single solar array
Mass: 227 kg (500 lb)
Hall-effect thrusters using krypton as the reaction mass, for position adjustment on orbit, altitude maintenance and deorbit

SpaceX plans 24 Starlink launches in 2020 and could have 4 launches in 2019.

This would mean that by the end of 2020, SpaceX would have about 1600 Starlink satellites in orbit. SpaceX could lose 80 (5%).

SpaceX plans to have orbital launches of the SpaceX Super Heavy Starship in 2020 and could have a manned Super Heavy Starship launch within 6 months of the first successful unmanned launch. A NASA environmental assessment indicates that SpaceX has filed a plan for up to 24 Super Heavy Starship launches.

In 2021, there could be 12 launches of Super Heavy Starship used for Starlink deployment. These would launch 3 to 4 times as many Starlink satellites. 12 launches of 180 Starlink satellites and 24 launches of 60 satellites would mean 3600 Starlink satellites in orbit by the end of 2021. This would mean nearly the entire 4425 first network would be in orbit. Two more Super Heavy Starship launches would complete the 4425 network by the end of 2021 even with some satellite losses.

In 2021 and 2023, SpaceX could have its planned 24 launches of Super Heavy Starship. This would be another 4,320 Starlink satellites each year. Some Super Heavy Starship launches would be used for other customers like the military or NASA. If we assume that Falcon 9 and Falcon Heavy launches fill in the gaps for the estimate. This means by the end of 2023, the second 7,518 Starlink satellites would be placed.

In 2024 to 2027, I will assume that SpaceX increases to 48 Starlink launches per year. Overall SpaceX would triple the launches but extra launch capability would be for other customers and projects. This would mean 8,620 satellites placed each year. SpaceX would have launch a refresh of the original 12,000 Starlink satellites, since they would only last about 5 years. Over 4 years, 34,480 Starlink satellites would be launched. This would be enough for 26,000 of the version 3.0 network and replacing 8,000 of the 1.0 and 2.0 networks.

A slight increase in launch cadence in 2026 and 2027 or a slight improvement in packing satellites would enable the full 30,000 satellites in the 3.0 SpaceX Starlink network to be placed by 2027.

SpaceX has discussed making a Starship 2.0 which would be twice the diameter of the Starship 1.0. If it was 50% taller, then it would be able to launch 6 times as many Starlink satellites. The larger Super Heavy Starship could start operation in 2028. This would deploy 1,080 Starlink satellites for each launch.

This would enable a Starlink 4.0 network with 120,000 satellites by 2030.

Starlink Business

SpaceX and Elon Musk will be made financially secure by 2023 and will have the $20 billion per year budget of NASA. If Elon has a 30X on his 54% share of SpaceX, then with Elon would have 30 times $10 billion in 2024 (50% of $20 billion in 2024). This means Elon would be worth over $300 billion without including any valuation for Tesla.

If Tesla still had any financial issues, Elon would be able to lend money from SpaceX to Tesla by late 2020 or 2021. Elon used Tesla to buyout Solarcity. In 2018, financial analysts speculated that Elon could his SpaceX stake as collateral in a buyout of Tesla. If SpaceX is worth $100 billion late in 2020 and then $200 billion in 2021, Elon would easily be able to fund a Tesla buyout with his $54 billion and then $108 billion of SpaceX (versus about $15 billion today).

In capital markets, low latency is the use of algorithmic (programmed) trading to react to market events faster than the competition to increase the profitability of trades. In 2007 a large global investment bank has stated that every millisecond lost results in $100 million per year in lost opportunity.

Laser light communication in a vacuum is physically 45% faster than communication through a fiber.

SpaceX will start generating substantial revenue in 2020 equal or slightly exceeding launch revenue. This was based upon 2017 SpaceX revenue projections from a 2017 Wall Street Journal article.

SOURCES – SpaceX, FCC, ITU, wikipedia, Wall Street Journal
Written By Brian Wang, Nextbigfuture.com

43 thoughts on “Future of SpaceX Starlink and Starships”

  1. 1 kW/m^2 is normal insolation on a clear day at noon at sea level. Hardly lethal.

    (edit: On 2nd thought, a kW of microwaves may be more dangerous than a kW of sunlight, but it’s still spread over a square meter. It’s not like if you’re standing next to a kW microwave bulb. That’d be a lot more concentrated.)

  2. ask yourself, what would happen if he completely dissmissed enviromentalism as a reason for solar power and electric cars?

    The press and activists would rip him a new one, that’s what.

  3. At 1GW of power and 1 km2 rectenna field, the intensity would be 1 kW/m2, i.e. clearly lethal. But you just have to seal of the area. The only hazard is if the beam drifts off.. Could be a spectacular deathray..

    On the upside, it would provide fodder for really good films. The hero has to collect a child from a rectenna field, and the beam is of. Will he make it out before it is turned on again..?

  4. He’s always been willing to rashly (and sometimes illegally) talk Tesla’s business case(s) up, but never down. Even for a rookie public CEO, that’s pretty straightforward.

  5. It’s probably the same reason he pushes car-sized tunnels instead of tunnels for smaller, cheaper PRT pods.

  6. Maybe – but he was saying this even back when he seemed to be saying whatever he wanted about Tesla…

  7. It would also be very interesting for the Moon. With the lousy capacity factors available on the Moon (<50% unless one of these Peaks of Not-Quite Perpetual Light actually pans out, and is located in an interesting place for a base), a high-power lunar base would require both at least 2x the nameplate capacity as the base power requirements, and massive amounts of storage.

    In contrast, an SPS in L1 or L2, transmitting in the 20-30 GHz range, could result in a much smaller landed cost for a rectenna, and a not-too-terrible cost for PV panels in orbit.

    L1 and L2 aren’t great from a distance standpoint: they’re 61,000 km from the surface, so diffraction losses are going to be non-trivial. But you can put a lot of extra PV in orbit for the cost of the batteries/fuel cells you’d need on the surface.

    It’s also a nice pilot plant for getting experience for SBSP on Earth: 10 MW to the lunar surface is a lot of power, but even if you needed to put 10x that amount of PV at L1 or L2, at 300 W/kg, that would be only about ~4 Starship-loads of stuff. At about 8 tankers per sortie and $20M per launch, you’d be looking at $720M in launch costs.

    I have no clue what a rectenna would wind up weighing, but it’s basically a few square km of conductive foil with some rectifiers sprinkled around. I’m guessing that it’ll be considerably less massive than the PV and batteries that it would replace.

  8. Musk has a fairly serious fiduciary conflict if he were to advocate for SBSP: He’s the CEO of a publicly-traded company that makes terrestrial solar power systems and battery storage. My guess is that the mild shade he’s thrown on SBSP may be half-hearted, but he’s really not in a position to advocate for SBSP as long as he’s Tesla’s CEO.

    Tesla could certainly decide to enter the SBSP business with the assumption that it might ultimately generate more money than their current power business model, but the transition would almost certainly hurt Tesla’s value in the medium term. It’s not a decision to be taken–or discussed–lightly.

  9. You’ve misunderstood. It isn’t the insolation that’s important, it’s that SBSP can have a >95% capacity factor (akin to nuke capacity factors), while terrestrial solar is stuck at about 28%–even in California, where most of the generation is in a desert in the extreme southern part of the US.

    SBSP would in fact be concentrated into big plants, just like you want. Rectenna farms are cheaper the bigger they are.

    The big fly in the ointment for SBSP is launch costs and launch cadence. But that assumes that you’re launching the materials for the SPS from Earth, instead of from the Moon or an asteroid. If you’re willing to commit to building out the ISRU and industrial capacity on the Moon (which is indeed a big “if”), then the cost would plummet, and the launch costs and cadence would be a minor part of the equation.

    I’m not at all sure that SBSP will be either economically or technologically viable: It requires a serious commitment to lunar resources and a willingness to work the kinks out of the technology, some of which may be show-stoppers. But I’d put it at about the same level of technological, economic and political difficulty as Gen IV nukes, fusion, widespread geothermal, or even storage that’s sufficiently cheap to enable 100% wind/solar net generation. In other words, we ought to look at SBSP pretty hard.

  10. Please re-read my post and search for the point where I advocate for space solar power. (Hint: it isn’t there.) Then try to examine your own biases and see where your mistaken impression that I am a solar or space-solar zealot came from. (Hint: I’m not.)

    Re: the grid – nuclear plants are also often built far from cities. The cost of land is trivial vs the cost of PV for a solar farm.

    SSP would have 1.5x higher peak intensity and near-24hr sun-tracking vs ~4hrs equivalent of peak intensity – ~8.5x not 2.5x.

    Concentrated PV with lightweight mirrors and active cooling should give at least another 5x improvement per unit energy-generating-system mass.

    So after 50% conversion losses, SSP is potentially ~20x more efficient than Earth PV. At Starship launch costs, that part might make economic sense.

    The big problem with SSP is that at ‘human safe’ beam power, the Earth-side collector could only collect a fraction of the energy per area of a solar farm – 30%-60% after factoring in continuous collection. For a cost-effective Earth receiver, beam power levels would be at least somewhat hazardous.

    In short – SSP might be viable with Starship launch costs – but the first unit (design) costs will be quite large, and environmental objections might kill it. The latter may be what actually puts Musk off SSP.

  11. Space solar power would work well for a Mars colony, though; Synchronous orbit is much lower, and a rectenna works much better covered in dust than a solar panel does. And in orbit thin film reflectors can easily concentrate sunlight to as high as you would want to minimize the actual panel mass.

    “massive infrastructure both in orbit and on the ground.”

    The infrastructure in orbit is pretty massive by current standards. On the ground? I wouldn’t call a rectenna field all that massive, and it’s transparent enough that you can grow crops under it, so the land doesn’t go unused.

  12. I’d be a bit worried about going too far in the direction of stealthing them. Might be nice to be able to find one of these even if it goes dead for some reason, and if it gets blown apart by space debris, the last thing you want is for the pieces to be invisibly dark.

  13. “So, if Musk has to choose between access to the chinese market and freedom of speech, liberties go out the window…”

    Not necessarily. If Musk cashes out his shares in Tesla then he won’t really care about China. If Starlink brings in $20 billion a year in revenue to SpaceX then he won’t care about Tesla and by extension wont care about China.

    I for one desperately hope that Mr Musk sells his shares in Tesla, strikes it rich with Starlink and then pours that money in to colonizing the Moon and Mars which would be quite the growth industry. Be the man who has a privately owned company that is insanely profitable and colonizing the solar system.

  14. Nutty environmentalists are of course still a problem, but the launch costs should basically by solved with BFR. It seem that you can do solar power at 1 kW per kg, which would put the launch cost at 270 USD per kW of power. Compare that to the cost per installed kW of the power source with the lowest up-front cost (gas) which is at least a 1000 USD per kW of installed power [1]. Wind power costs about 1600 USD per kW, and with a 20% availability results in an effective cost of 8000 USD per kW… In short, the launch costs should be negligible compared to costs of other power sources with the advent of the BFR system.

    Of course, other parts of the system may also be very heavy such as the microwave transmitter and rectenna…

    (1) https://en.wikipedia.org/wiki/Cost_of_electricity_by_source

  15. The problem is getting the power to the ground. They looked at the idea in the ’80s, and the best idea was a microwave beam down to a large rectenna farm in an unoccupied area.

    The environmentalists went bugf*ck. “YOU WILL FRY FLOCKS OF BIRDS! MICROWAVES! OMG!!11!!”

    And launch costs were ridiculous. This was post-Saturn V and just barely into the Shuttle era. It would have been way too expensive to put up the needed satellites with Titans or Deltas. And we won’t even talk about the exhaust byproducts.

    So the idea went nowhere. And it’ll probably go nowhere in the future – because ground-based power is safer and easier to maintain. (Which is a damned shame, because the idea is a very interesting and appealing one, but… it’s a lot like HSR. Not economically feasible, no matter how tasty it may be.)

  16. It’d be cool if every Starlink had a telephoto-lens ultra-res camera on it, offering very up-to-date Earth imaging. It probably be fixed orientation though – otherwise would induce movements that could interfere with laser communications between sats.

  17. Musk has said space solar power is nonsense.

    However the reason he gave (conversion inefficiencies cancel the value of stronger sunlight in orbit) didn’t account for typically having 24hr/day sunshine, unlike Earth-based solar.

    Maybe he’ll reconsider once Starship is flying and needs more payloads.

  18. Though it smells like SpaceX is doing an application documentation steamroll job on the ITU, hoping their bureaucracy can’t handle the quantity fast enough to reject before he gets his minimum viable constellation up, creating a defacto situation regarding spectrum usage and by default following FCC rules.

  19. Why is Elon not branching out to space based solar power? I know that the starlink may be more profitable, but space based solar power could really impact the environment in a positive way. And it right up the alley for what Elon Musk does well: mass production of units, innovation to get the weight down to 1 kg per kW coupled with (possibly) lithium batteries to smoothen out power generation over 24 hours (may not be needed, but is the upper time limit on smoothing needed for space based solar).

    After a quick googling, it seems that the BFR could place cargo at geosynchronous orbit at a mere ~270 USD per kilogram, which is far below the threshold where space solar could become viable. So what is stopping him?

  20. I have my doubts. Did you see how NBA bent over backwards for China? Apparently, it’s quite brave to denounce cops as “pigs”, but blandly supporting freedom in Hong Kong is “culturally insensitive”. Its part of the sacred chinese culture to be under communist dictatorship….

    So, if Musk has to choose between access to the chinese market and freedom of speech, liberties go out the window…

  21. Nobody ever bothered to try to lower albedo on communications sats before. There’s nothing on them that has to appear reflective from the ground. The PV panels would be sun facing. Everything ground facing could be coated/masked with something like vantablack and be effectively invisible even pretty closeup. A sky with thousand of satellites in it would look no different to the human eye and be little different to telescopes.

  22. At some point, probably long before 30,000 much less 120,000, Starlink would scale up the size of the individual satellites rather than enlarging the Constellation. For example with the same flat packing in Starship there could be 4 current form sats attached in a layer with more than 60 in the stack.

  23. I can’t find any information on the actual size of these satellites, but some earlier speculation was 1.5×1.5 m, so a total area of 2.25 sq.m.

    With a million in orbit, let’s say 1/3 at any time would be shadowing the earth.
    That’s 333 333 x 2.25 = 750 000 sq. m.

    Total projected area radius of the earth = 127.5 trillion sq. m so the reduction in sunlight would be 1/170 million…. not enough to cool the Earth significantly.

    Oh well. I thought he was going to pull another rabbit out of his hat.

  24. he International Telecommunication Union (ITU) (French: Union Internationale des Télécommunications (UIT)), originally the International Telegraph Union (French: Union Télégraphique Internationale), is a specialized agency of the United Nations that is responsible for issues that concern information and communication

  25. Maybe Elon could buy all the wireless carriers in North America and create a continent wide, public, self funding, wireless infrastructure project.

  26. Game on.
    As I have previously stated, the signal should be made available to all repressed peoples. Can’t stop the signal.
    Plus- what a perfect cover for LEO MilCams. Which one do you shoot down? Don’t be surprised if we find out one day that Musk had help with these authorizations.
    Neo-Hughes LOL.

  27. The International Communications Union is an American agency? That’s news to me!

    The FCC approves the use of bandwidth when communicating with US targets; other countries will require separate permissions to access the network.

  28. What is the value of Starling as a weapon system? Quite high if multiple days could hit within a few meters of any point on earth. Maybe the true value is in extortion.

  29. Elon did say that SpaceX would drop the albedo on the Starlink satellites.

    Plus the number of satellites will add up. 4425 + 7518+ 30,000 + 120,000 = 162,000

    Also, 4.0 could be more if they try to max out the Starship 2.0 with 6X the capacity. 6X capacity and 3X the flights in a year. There about be 300,000 satellites per year or for 1 million from just over 3 years of launches. That could be a 5.0 network. Almost one satellite replacing each cellular ground station. The overall system bandwidth. Each satellite is currently 20 Gbps. In ten years each satellite could be 20 Tbps. In 10 years with 1 million satellites the overall system capacity could be 20 million terabits per second.

  30. I find it odd that an American agency has the jurisdiction to fill out the global near space with American Companies satellites without checking the impact of obstructing the skies on humans and on the globe. Anything in our natural environment has an importance, even if the dogma has never considered it.

  31. which coincidentally, the very same internet constellations will allow, by capitalizing SpaceX to the point Starship and moon projects will get cheap enough.

  32. 120,000 satellites by 2030

    Oh boy, if critics were already whining about the end of astronomy and Kessler syndrome, I can’t imagine how they will be reacting by 2030.

    The number of shooting stars resulting from old satellites doing re-entry would be probably epic too.

    And at those scales of launch and cost, it starts to make sense to place the cloud infrastructure literally above the clouds.

    We can imagine stacks over stacks of off the shelf server satellites, fed with Sun’s light and designed for an operational life of a few years, and profiting from the low ping latency of a transponder just a few hundred miles over the vacuum above our heads, connected in a mesh of many others.

    By then, your computer can be in space and have gone much further than you or any of its users just in miles traveled.

Comments are closed.