Amazon LEO (formerly Project Kuiper) Broadband Constellation

[leeloodallasmultipass]

NOT; build an antenna terrestrial is a lot cheaper than launch thousands of sats.

It can't compete vs terrestrial.

What do you mean by this? You cannot replace thousands of satellites with a terrestrial antenna. Those satellites replace hundreds of thousands of cell towers and/or millions of miles of copper and fiber, so I assume you mean something else.

No one replaces fiber though(Satellites uses fiber too). But I think he means it is much easier to put some 5g cell tower(with fiber connections) and provide service to a community than put thousands of satellite up there, which I agree. At no point establishing leo satellite isp business cheaper than wireless service. Yes, big companies can shoulder the bill for some time but economically wireless is so much cheaper, it is not reasonable to try and compete with them via satellite internet(any kind of satellite).

Basically anyone can be wireless provider but leo satellite market is multibillion dollar intiative.

[leeloodallasmultipass]

Spacecrow, Amazon leo's standart(pro) terminal is similar size to starlink mini with better performance(on the paper). Nano is about portability. Nano terminal is much more portable than mini, thats what I am saying.

Dishytech channel has  a video about that comparison. 40(not 30) percent area is about right.  11.75 x 10.20 vs 7x7... You can do the math. Nano terminal itself without attachments is about kindle weight. Amazon says "Weighing just 1 pound"

Absolutely nobody is comparing them by area when deciding which to buy. Both are small enough to fit in a backpack, but too large to fit in a pocket, and thats about as far as the dimensional comparison is going to go.

Weight is most important for high portability, and Amazon says it weighs 2.2 lb with the stand. Maybe it weighs 1 lb without it, but nobody has actually weighed one, and 1.2 lb seems awful heavy for a little tripod.

Leo's Pro terminal is 5.3 lb, more than twice as heavy as Mini. Nobody is going to be carrying that anywhere, so they aren't really competing on portability.

"Absolutely nobody is comparing them by area when deciding which to buy." One of them is not even on the market, give people some time. And of course people will look at the dimensions when they want to buy a portable terminal. Would you look at the color and texture instead(some people would though)? Smaller the better and one of them is around 40 percent of the other.

"but nobody has actually weighed one" except Amazon of course.

"Leo's Pro terminal is 5.3 lb, more than twice as heavy as Mini." mini has no big attachment whereas leo pro has big attachment for roof installment. Remove that and you will have similar sized antennas. Nonetheless, nano is the portable one not pro and nano is very promising at that front(if it performs).

[launchwatcher]

eventually they'll be able to challenge fiber too.

A head-to-head challenge to fiber is a tall order.   Performance-wise, current fiber is not standing still.   Current buildouts are typically offering XGS-PON which offers 10Gbit/s symmetric with ~5ms latency -- my 10G service (for which I'm paying $50/month) shows ~8Gbit/s throughput on speed tests.  There are even faster standards in the pipeline -- 25G and 50G -- which will in most cases work over already-deployed fiber outside plant just by swapping the head-end gear and the ONT box at the customer.   And for interactive/gaming use cases, even LEO satellites will have trouble with latency relative to fiber. 

IMHO, 1Gbit/s is more than sufficient for my needs -- I only have the 10G service because it's a one-size-fits-all deal from the ISP; the value of faster PON standards is more in the overall capacity they create within the network.

Fiber's biggest weakness is the often-Herculean effort required to get the outside plant built with 100% coverage.   Anything from obstructionist city councils to rotten utility poles to overfull underground conduits can stall deployment for years.   IMHO the best competitive angle long term is to go after fiber's gaps rather than head-to-head.   In areas where fiber has coverage, satellite can still get customers who need intermittent mobility (RV's) , the budget minded (lower bandwidth for less than what the fiber ISP's charge), and/or occasional backup service for when the fiber ISP has a bad day.

[Tywin]

eventually they'll be able to challenge fiber too.

A head-to-head challenge to fiber is a tall order.   Performance-wise, current fiber is not standing still.   Current buildouts are typically offering XGS-PON which offers 10Gbit/s symmetric with ~5ms latency -- my 10G service (for which I'm paying $50/month) shows ~8Gbit/s throughput on speed tests.  There are even faster standards in the pipeline -- 25G and 50G -- which will in most cases work over already-deployed fiber outside plant just by swapping the head-end gear and the ONT box at the customer.   And for interactive/gaming use cases, even LEO satellites will have trouble with latency relative to fiber. 

IMHO, 1Gbit/s is more than sufficient for my needs -- I only have the 10G service because it's a one-size-fits-all deal from the ISP; the value of faster PON standards is more in the overall capacity they create within the network.

Fiber's biggest weakness is the often-Herculean effort required to get the outside plant built with 100% coverage.   Anything from obstructionist city councils to rotten utility poles to overfull underground conduits can stall deployment for years.   IMHO the best competitive angle long term is to go after fiber's gaps rather than head-to-head.   In areas where fiber has coverage, satellite can still get customers who need intermittent mobility (RV's) , the budget minded (lower bandwidth for less than what the fiber ISP's charge), and/or occasional backup service for when the fiber ISP has a bad day.

Exactly.

[DanClemmensen]

NOT; build an antenna terrestrial is a lot cheaper than launch thousands of sats.

It can't compete vs terrestrial.

What do you mean by this? You cannot replace thousands of satellites with a terrestrial antenna. Those satellites replace hundreds of thousands of cell towers and/or millions of miles of copper and fiber, so I assume you mean something else.

No one replaces fiber though(Satellites uses fiber too). But I think he means it is much easier to put some 5g cell tower(with fiber connections) and provide service to a community than put thousands of satellite up there, which I agree. At no point establishing leo satellite isp business cheaper than wireless service. Yes, big companies can shoulder the bill for some time but economically wireless is so much cheaper, it is not reasonable to try and compete with them via satellite internet(any kind of satellite).

Basically anyone can be wireless provider but leo satellite market is multibillion dollar intiative.

One cell tower serves one cell. A megaconstellation serves hundreds of thousands of cells.

You are confusing two kinds of fiber. All big ISPs, including Starlink, use a fiber backbone. The satellites do not (mostly) replace this. However, a constellation does replace all the "last mile" fiber or copper, which probably averages more than 0.1 mile per customer and which is expensive to install and maintain.

[envy887]

Satellite can definitely beat fiber at very low customer population densities, while it definitely cannot beat fiber at very high population densities. Tech improvements for each can shift the exact population density where they are equal, but won't allow one to wipe out the other.

When talking about satellite challenging or competing with fiber, most people are referring to that segment of population density where there's no clear winner. Not dense apartment blocks or even in subdivisions or small towns, and also not extremely rural areas where the "last mile" connection may actually be multiple miles of dedicated line per customer. Most people are also not talking about satellite serving an outright majority of fixed broadband connections, but about the 10 or maybe 20% of customers that are rural, but not so rural that satellite is a clear winner. That's the front of head-to-head competition where the two technologies vie for customers.

[leeloodallasmultipass]

I read somewhere the Ariane6 launch is early 2026(my guess february) and will carry 32 Leo satellites. It is proportionate Atlas V and Ariane 6 payload capacity.

[DanClemmensen]

I read somewhere the Ariane6 launch is early 2026(my guess february) and will carry 32 Leo satellites. It is proportionate Atlas V and Ariane 6 payload capacity.

Doing a straight proportion, I get a number of about 31. However, it's more complicated than that. Assuming Amazon uses the same type of dispenser (multiple rings of 9) then adding the fourth ring adds another 9 slots, and maxes out at 36. A partially-full ring will have a slightly higher extra mass per satellite. The third ring on F9 was partially full (six of nine) to get to 24 satellites. I do not know if symmetry matters. If so, then 30 or 33 (three of nine or six of nine) would be preferred.

[GewoonLukas_]

I read somewhere the Ariane6 launch is early 2026(my guess february) and will carry 32 Leo satellites. It is proportionate Atlas V and Ariane 6 payload capacity.

Doing a straight proportion, I get a number of about 31. However, it's more complicated than that. Assuming Amazon uses the same type of dispenser (multiple rings of 9) then adding the fourth ring adds another 9 slots, and maxes out at 36. A partially-full ring will have a slightly higher extra mass per satellite. The third ring on F9 was partially full (six of nine) to get to 24 satellites. I do not know if symmetry matters. If so, then 30 or 33 (three of nine or six of nine) would be preferred.

They have smaller dispenser rings with less satellites. The launches on Falcon 9 didn't have any empty slots.

[leeloodallasmultipass]

https://www.arianespace.com/news/canopee-sets-sail-for-ariane-6-flight-va267/

Early 2026, 32 satellites.

If I remember correctly, 2 p120c booster version a64, so I expect a relatively long period between flight 2 and 3. New, more powerful version will probably lift 35-37.

Also, that looks cold.

[russianhalo117]

https://www.arianespace.com/news/canopee-sets-sail-for-ariane-6-flight-va267/

Early 2026, 32 satellites.

If I remember correctly, 2 p120c booster version a64, so I expect a relatively long period between flight 2 and 3. New, more powerful version will probably lift 35-37.

Also, that looks cold.

Reportedly upto 40 sats per launch in the contract by the time all Block-2 upgrades are phased in.

[leeloodallasmultipass]

https://dplnews.com/directv-cuando-llegara-internet-satelital-amazon-leo-argentina-chile-y-brasil/

This source says "July 2026 in Argentina, starting with Patagonia; October in Chile; and at the end of the year in southern Brazil." Also Amazon will sell directly to customers in Brazil but not all Directv agreement zone.

Proper hassle free service will come only in 2027 in my opinion. Then launch bottleneck  would be solved too.

[XRZ.YZ]

American Airlines Group Inc. has held discussions with Amazon.com Inc. about using the Amazon Leo satellite-based internet service for in-flight Wi-Fi, part of a campaign to lure premium customers and better compete with US rivals.

Even as carriers such as United Airlines Holdings Inc. have signed deals with Elon Musk’s SpaceX to use its Starlink network of low-Earth orbit satellites, American is keeping its options open, Chief Executive Officer Robert Isom said in an interview.

“While there’s Starlink, there are other low-Earth-orbit satellite opportunities that we can look at,” he said.

American has talked with Amazon, Isom added, but he declined to comment on the status of any discussions.

“We’re making sure that American is going to have what our customers need,” he said.

A representative for Amazon didn’t immediately respond to a request for comment.

Amazon has more than 150 satellites in space and aims to build a network of more than 3,200 units. The service, formerly called Project Kuiper, has contracts with customers like JetBlue Airways Corp. and L3Harris Technologies Inc., according to the company.

American currently works with satellite companies such as Viasat Inc. to provide connectivity to passengers and is focusing on offering free Wi-Fi to members of its loyalty program, starting in January, through a partnership with AT&T Inc.

As satellite technology improves, American will have opportunities to work with more providers, Isom said.

“I do think that there’s going to be advancements and, as that comes, I think that there’s going to be better deals to be had,” he said.

In-flight internet access has long been an unreliable and expensive offering for passengers, but many carriers are striving to improve their service thanks to upgraded satellite-based connectivity.

Starlink has signed carriers including United, Alaska Air Group Inc. and Emirates, Bloomberg has reported.

https://www.bloomberg.com/news/articles/2025-12-10/american-airlines-in-talks-with-amazon-for-leo-in-flight-wi-fi?srnd=homepage-americas

[leeloodallasmultipass]

Amazon will eventually get into airplane market but they may not be very competitive before polar launches. Starlink is going to be better at this market for a couple more years(upto 2028). Even oneweb has something to offer but not Amazon. That is why Amazon is going after the oems at this stage not the customers themselves, in my opinion.

[Tywin]

Could Amazon buy spectrum from Globalstar or Iridium?, like SpaceX did with Echostar...

I the long term, the D2D is a good extra market...

[leeloodallasmultipass]

Could Amazon buy spectrum from Globalstar or Iridium?, like SpaceX did with Echostar...

I the long term, the D2D is a good extra market...

Don't think so. But they might support Ast mobile with a minority partnership. Amazon wants to be platform of platforms as I umderstand and too much individual  customer facing ventures dont align with (my perceived)  strategy of Amazon. But who knows?

[sstli2]

They currently have 153 satellites, with two of them being deorbited, so 151.

- LA-04 next week will add 27 satellites.
- LV-01 (NET Jan 2026) will add 45 satellites.
- LE-01 (NET Jan 2026) will add 32 satellites.

So roughly 255 satellites. To begin initial service is 578 satellites, or 323 satellites to go. That's ~7 Vulcan launches, ~10 Ariane 6 launches, ~12 Atlas V launches, or ~6 New Glenn launches. Q4 2026 is probably where this shakes out.

[catdlr]

https://twitter.com/Amazonleo/status/2000694865948966980

Amazon Leo

@Amazonleo
·
Back on the pad: Atlas V is loaded with another batch of Amazon Leo satellites. Liftoff of LA-04 — our fourth mission on a @ulalaunch Atlas V and seventh of the year — is set for tomorrow, Dec. 16, between 3:28 and 3:57 a.m. ET.

Follow the launch coverage on NSF here: https://forum.nasaspaceflight.com/index.php?topic=63924.msg2743147#msg2743147

[edzieba]

It's not clear that Amazon has any intention at all of marketing to retail customers with their minimum constellation of 578 satellites. And because commercial contracts are negotiated, I wouldn't hold your breath on any kind of standardized pricing becoming available anytime soon.

Come back in 2027.

In addition, service start can be in advance of full 24/7 coverage, for customers who do not require that. See: Starlink's initial service offering.

Starlink had continuous coverage when they started, even for the invite-only beta service.

Before that it was non-continuous, but only accessible to employees and investors.

The "Better than Nothing Beta" was open to the public, and was still during the initial buildout phase where connection interruptions were commonplace due to 'holes' in the shells - as of 2020 when the beta became available, about 1/2 of the V1 satellites were launched (but many were still 'phasing' to their slots so were not in their final positions and generally not providing service due to being oriented for drag control) and <10% of the current constellation size. In the initial phases of the beta, connection dropouts occurred several times per hour.

[sstli2]

Is there any significance to the naming of the Kuiper satellites on Celestrak's supplemental GP? https://celestrak.org/NORAD/elements/supplemental/table.php?FILE=kuiper&FORMAT=tle

The constellation started at Kuiper-008, but since then it has consistently incremented by the satellite count, across launches.

This most recent launch, LA-04, is different. It starts at Kuiper-179 and goes all the way up to Kuiper-265, skipping various numbers in between, covering a range far greater than the 27 satellites that were actually launched.

Does this imply that they took incoming shipments (probably sequentially numbered) and separated them out across multiple concurrent encapsulations, and that we will see these numbers fill in once LE-01 and LV-01 are launched?

If that's the case, then that would further imply there's 265 - 179 - 27 or at least 59 satellites manufactured and waiting. That's not necessarily a surprise; we already know of the 32 satellites that were shipped for LE-01 on Ariane 6. I just thought it was interesting that the satellite names seemed to imply of the existence of these satellites.