Once upon a time, back in the 1990s, there was a Thing where everyone got super excited about constellations of satellites in LEO (low earth orbit) that could support communications.
https://en.wikipedia.org/wiki/Iridium_satellite_constellationAt that point in time, it was basically a voice only operation, AFAIK. Heady stuff, very niche. Launch costs turned out to make this prohibitively expensive.
A Very Few Years Later, Elon Musk wanted to buy a rocket, did not succeed and decided, As One Does, to make his own. And thus was SpaceX born. A few years later, Elon Musk was already forcing change on the federal government and moving things in the direction of commercial space development vs. the old-skool procurement system of bidding and contracts. There were some spectacular failures (and not just with SpaceX) and then things turned around. The rockets were launching and later landing for reuse, and Musk was collecting on payments from the government, and Musk announced Starlink. I _think_ this happened after Musk and Greg Wyler of OneWeb (or whatever) parted ways.
Starlink is another constellation of satellites, a really big one, but it’s not a voice only thing and it is not niche. Its first batch launched in 2019, and when Russia invaded Ukraine again, Starlink was in the news as an important component in sustaining Ukrainian communications as they resisted. The broadband everywhere feature also got a big bump-up in interest from the general public as a result of pandemic era movement out of cities and to rural areas.
Shortly after Starlink started launching satellites, Amazon announced Project Kuiper. While pandemic era restrictions and supply chain issues and delays associated with various launch vehicles resulted in delays in Project Kuiper, by the end of 2023, the first two prototype satellites had been launched on a ULA Atlas V, and testing was successful.
In 2021, Google announced that it would be providing cloud and connectivity to Starlink:
https://www.cnbc.com/2021/05/13/google-cloud-wins-spacex-deal-for-starlink-internet-connectivity.htmlMicrosoft’s Azure also has some kind of Starlink connectivity as well.
Microsoft, Google, Meta/Facebook and Amazon have for years individually and collaboratively been building private fiber networks connecting their datacenters, connections which extend across the ocean and dwarf the fiber networks of the world’s telcos.
https://www.wsj.com/articles/google-amazon-meta-and-microsoft-weave-a-fiber-optic-web-of-power-11642222824The conflict in Ukraine spotlights the value of Starlink and potential competitors such as Project Kuiper, to communities that are trying to survive and rebuild in war-torn areas, and to anyone attempting to defend an area militarily. The pandemic itself and movement out of coastal cities to more rural areas highlights the value of Starlink and potential competitors such as Project Kuiper to individual consumers who struggle to access broadband in places they would like to live.
Cruiseships and trains are making use of Starlink, in an effort to improve wifi service to their customers:
https://www.wsj.com/articles/train-wi-fi-can-be-terrible-operators-are-looking-high-and-low-for-solutions-c245a592Hawaiian, Qatar and New Zealand airlines are all offering Starlink based wifi on at least some of their flights.
I blogged extensively recently about a forum thread about Project Kuiper. The thread as a whole is aggressively hostile to Project Kuiper, and really any constellation which is not Starlink. Because it exists in a forum context focused on space, the expertise in the community is more in launch related areas than it is … anywhere else. Their capacity to make sense of the purpose or even the realpolitik aspects of regulations is extremely limited if not non-existent. They don’t actually understand anything I mentioned above about the pipes that the internet runs through these days (as near as I can tell, they don’t know that Microsoft/Alphabet/Meta/Amazon own any fiber at all). They don’t have any real awareness of the business models of any of those companies, nor do they have any apparent understanding of the relative scale of the companies involved. They have no apparent understanding of either corporate strategy or military/national security strategy. They are aware of the term “first mover” but do not appear to have any understanding that, you know, first mover doesn’t always “win”. Their perspective on a shareholder lawsuit (by a union pension fund) against Amazon and individuals at Amazon, combined with their perspective on deadlines from the FCC, the ITU and similar, suggest an almost bogey-man like response to regulators that is very much at odds with how Musk has behaved with respect to regulators.
It makes for an interesting thread that spans the 4ish years since Project Kuiper was first announced until now.
I learned a ton — notably, I didn’t know about the pipes either! — by reading this thread and thinking through the screwy arguments in the thread, and researching the regulations and the history of constellations and so forth. There are aspects to the debate that we really won’t know until we know. I found this thread because I wanted to get some sense of the size of the market for Project Kuiper, and this thread did not help me figure that out, except in a “negative space” sense. I’m very sure that the market for Starlink and Project Kuiper and similar constellations is much larger, and very, very different from what the participants in this thread describe as their imagined market(s).
One major argument is that SpaceX has an uncloseable lead in innovating in launch vehicle development. I find this argument pretty unlikely.
https://www.reuters.com/technology/space/china-launches-powerful-jielong-3-rocket-paves-way-more-commercial-missions-2024-02-03/Among other things, China has been launching from barges rockets with larger payloads than the US has managed to do.
A second major argument, related to the first, is that Project Kuiper will suffer from having to buy its launches in the marketplace, vs. the integrated launches available to Starlink as a part of SpaceX. The thread in general argues that Amazon should purchase launches from SpaceX as the cheapest/most reliable launch vehicle versus using multiple launch partners. They are particularly antagonistic toward Blue Origin, which is perceived as using staff cast off by Musk, and being slow. There were a range of reactions to Amazon adding Falcon launches to its contracts (one person believes that this is directly in response to the shareholder lawsuit; in general, posters think they will need many more Falcon launches and this is too little too late to meet the FCC launch deadlines).
The launch supply argument has several faces to it. First, it has been a source of delay. Second, it has been a cost driver. Repeatedly expressed in the thread is the idea that higher launch cost will inevitably lead to higher service costs. In conjunction with the belief that the market for Starlink, Project Kuiper etc is real but quite limited, the conclusion reached is that Project Kuiper will fail competitively because it will be a higher cost product. While there is some understanding that Amazon might be able to absorb some of the cost, in general the more long-winded, frequent posters assert that over time, the difference in cost structure will overwhelm Amazon’s ability to subsidize and Amazon will abandon the product.
This is sort of weird? I mean if the market is big enough that Amazon can’t subsidize, then isn’t it more than big enough to support more than one provider?
Another component of the launch supply cost argument involves future generations. LEO satellites have comparatively short orbital lifetimes, and will thus need to be replaced if the service they provide is to continue (5-7 years seems typical, altho within the thread a lifetimes such as 3 years are mentioned, altho without any support). This is perceived as a problem for Amazon’s Project Kuiper, as they will constantly be having to pay for expensive launches compared to SpaceX’s StarLink.
Let us for a moment consider whether launch cost is likely the dominant cost of this product. It’s not like the satellite itself is free, for example. For the service to be useful, however, there is a substantial amount of terrestrial support needed. Also, probably not free! Finally, there is integrating the satellite service, whatever it is imagined to be, into whatever is then offered to the customer. StarLink has a DTC sort of approach, as with many Musk projects. We don’t know what Project Kuiper will be offering, but it appears to be more enterprise oriented. The integration of Starlink into the google cloud and possibly also Microsoft Azure further suggests that in general, pricing of this particular component may be difficult to consider separate from the rest of the universe of services it is offered as a part of. Whenever someone sits down to try to figure out the input of labor vs. food costs in that fast food burger, they discover quite rapidly that branding costs are probably a bigger factor than anything else.
Finally, let us imagine a day 3-7 years in the future when Starlink, some non-Starlink constellation (probably Project Kuiper, but potentially something else like Rivada or whatever) are both available to customers such as airlines wanting to provide wifi on flights, or people who want broadband in the middle of nowhere, or some brave soldiers somewhere defending something from someone else. Let us imagine that Starlink has been the dominant provider from now until then, and the other provider is a relatively newer entrant trying to build a customer base. Do we believe that Starlink will be such a good internet service provider that no one will consider switching?
Because let me tell you, that would be a genuinely new feature of providers of internet service.
![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}