Hacker News Reader

Comment by pmg | Hacker News Reader

Feels like this is a direct response to the Amazon / Project Kuiper 7” terminal announced last year: https://www.aboutamazon.com/news/innovation-at-amazon/heres-...

View on HN

The major difference being the Starlink mini terminal has 6,078 satellites to connect to today to provide functioning internet, vs the Project Kuiper terminals that can connect to a grand total of zero satellites.

The Kuiper proto-network came online in November last year with two satellites [1][2]. If you live near their Denver office you can pick them up on Ka band or just look up their id’s on CelesTrak or NORAD (2023-154A and 2023-154B).

Having had one of the first 1,000 dishy V1’s (at least by MAC address) I’m as big of a fan of Starlink as anyone if not more so, but even I have some reservations about cheerleading a low-earth orbit communication monopoly. Yes these other guys are behind but it’s intellectually dishonest to call it vapor ware or “zero”.

I plan on supporting both networks and more if possible just as with GNSS.

[1] https://www.reuters.com/technology/amazons-prototype-kuiper-...

[2] https://youtu.be/K34ch_L_YxY

A direct response to vaporware?

This much more likely to be something that had been in the long term plan for many years.

Amazon is announcing stuff because they are spending 10e of billions and have nothing to show for it, and wont have anything for years. They need to show progress to the quarter to quarter investor.

Frankly the idea that Starlink is passivly reacting to Amazon annoucments is a bit silly. The difficult part is to actually produce these terminals at high volume cheaply, not releasing a few renders.

They have at least three Kuiper launches booked in mid-2025, that’s not far off! It’s awesome that they have the flexibility to choose different launch partners and can use the most reliable and flight proven technology.

https://www.aboutamazon.com/news/innovation-at-amazon/amazon...

> It’s awesome that they have the flexibility to choose different launch partners and can use the most reliable and flight proven technology.

Does Starlink not have that option? Do you think other launch providers would refuse to take SpaceX's money?

Also this isn't a situation where you need reliable and proven. If option A has 99.9% reliability and costs $80M to launch $20M of satellites, and option B has 95% reliability but costs $60M to launch $20M of satellites, then you should go with option B for your fleet.

> Does Starlink not have that option?

Starlink already uses "the most reliable and flight proven technology" (imho) so, yes, they have that option.

And yes, I should have added something about cost effective.

When you said "It’s awesome that they have the flexibility to choose different launch partners and can use the most reliable and flight proven technology." I thought you were listing that as an advantage they have compared to Starlink. If every competitor has the same flexibility, then why bring it up?

It was a tongue in cheek comment that was maybe a wee too subtle. To the extent that Amazon launching Kuiper is tied to (maybe still?) majority owner Bezos' interest in building supportive business for Blue Origin, then Kuiper launching on SpaceX is an irony that cannot go unobserved. One might even say Blue Origin as an orbital launch provider is more "vaporware" than Kuiper itself given the latter's test existence in orbit.

Starlink just had launch #173 and like 70 of them was only in 2023:

https://en.wikipedia.org/wiki/List_of_Starlink_and_Starshiel...

> It’s awesome that they have the flexibility to choose different launch partners and can use the most reliable and flight proven technology.

Except that likely is Falcon 9, and SpaceX gets launches at cost.

3 launches in mid-2025 isn't really all that promising an indicator for a constellation that needs to launch thousands of satellites. They'd need signs of an extreme ramp up in launches coming, particularly since so many of these earlier launches will have to be on external providers like ULA and ArianeSpace, who are already running delays, need to build entire rockets each time and have little experience with high launch cadences. At least whenever Blue Origin finally gets to flying New Glenn, they'll be able to avoid having to build a brand new booster each time.

The most reliable and flight proven launcher, F9, usually gets in a launch every 3 days and it's still going to have to keep that pace up until Starship is ready to dump hundreds of starlinks into orbit in one launch.

> a constellation that needs to launch thousands of satellites.

If I were the folks at Kuiper or any other would-be LEO RF communication constellation:

  * Starlink is proving the value prop to investors and consumers
  * This is an early stage this sector--there will be tech improvements
  * Improvements in satellites mean competitors can enter at vN+1
  * SpaceX/Starlink has some anti-trust risk--they might throw a bone to a weak "competitor" a la Apple/Microsoft 1997 [0]

0. https://youtu.be/WxOp5mBY9IY?t=44

Regarding the anti-trust risk, SpaceX has been very clear that they'll launch competitor's satellites just as fairly as they do any other payload. They only serve to benefit from it, as the profit from every commercial launch helps offset their costs on building out Starlink and on Starship R&D (which is also why Amazon didn't buy F9 for building Kuiper until New Glenn is ready).

IIRC there's even an ongoing lawsuit from some shareholders against Amazon for not acting in shareholder interest by paying for other much more expensive and less proven launch services rather than flying F9.

SpaceX satellites have a designed lifespan of approximately 5 years. They'll sustain their launch pace just to keep up the constellation. There's no real vN+1 advantage for competition.

> SpaceX satellites have a designed lifespan of approximately 5 years.

I wasn't aware of that lifespan but I knew it was finite. Let's say 6k sats at 20/launch is 300 launches to refresh the fleet over 5*12 months makes for 5 launches a month [0].

If the first v1's were launched November of 2019 then they will drop out of orbit later this year. SpaceX started launching the v2 minis about 9 months ago. Wikipedia claims 300/Starship. Let's round down to 200 to account for Elon-time and maybe launching a mix of Starship and F9. That would bring the needed operational cadence down to .5/month [1].

Could a well capitalized competitor start at vN+1 or vN+2, build out while Starlink air-ground combination is stalled at vN due to encumbancy, and make progress in high margin geographies proven out by Starlink already? I think it could look like the cell market with a few major players that either get squeezed out or created during any technology refresh.

Would it be more difficult because whatever that gets created is not a mutually supportive vertically integrated enterprise but fragmented? Sure.

0. 6000/20/5/12=5

1. 6000/200/5/12=.5

They indicated previously they changed their mind about using Falcon 9 and Starship.

By 'they' do you mean Amazon?

Yes. See https://spacenews.com/as-clock-ticks-on-amazons-constellatio...

Oh, interesting!

3 launches don't make a constellation.

> It’s awesome that they have the flexibility to choose different launch partners

In theory yes, in practice it isn't because they can't really use SpaceX (fully) and the other options aren't nearly as good.

> can use the most reliable and flight proven technology.

I'm not sure if you know anything about rockets. They bought some old Atlas 5s, but those are finished production. After that they have to buy unproven new designs.

And again, Falcon 9 is the most reliable and flight proven rocket in the world and they are simply not gone use Falcon 9 beyond the minimal.

> It’s awesome that they have the flexibility to choose different launch partners and can use the most reliable and flight proven technology

Heh, Amazon was going to forgo that for petty/corrupt reasons, but were sued by their shareholders.

https://arstechnica.com/space/2023/12/a-bitter-pill-amazon-c...

Capitalism at its finest. Tell that billionaire he can't blow all the vapor all the time.

A year away might as well be a decade in vaporware terms.

I mean Elon has mentioned the laptop sized dish for the past couple years.

Kuiper doesn't exist

Reply on news.ycombinator.com