Amazon Project Kuiper Broadband Constellation

[meekGee]

If they had chosen another established provider with no eol in sight, the argument might have been different.

There were no options for established launch vehicles with no EOL from western-friendly countries large enough to be cost effective other than Falcon. I don't think one should blame them for not picking a good option when there weren't any good options available.

I am sure they'll wiggle out of the minimal launch requirements, and they're behaving as if they think so too.

That complacency is reminding me of ULA.

This is what's killing them, not the regulations.

[DanClemmensen]

If they had chosen another established provider with no eol in sight, the argument might have been different.

There were no options for established launch vehicles with no EOL from western-friendly countries large enough to be cost effective other than Falcon. I don't think one should blame them for not picking a good option when there weren't any good options available.

A case can be made that Amazon made a defensible choice in April 2022, when they contracted for launches on Vulcan, Arianne 6, and New Glenn. However, their failure to adjust to reality since then is indefensible. At an absolute minimum, they should have contracted for a whole lot more than three F9s in December 2023, but realistically, this re-evaluation should have occurred no later than April 2023. Since they failed to do this, FCC should deny the extension.

The whole reason for these deadlines in the first place is to weed out applicants that cannot make timely and effective use of the licensed spectrum, which is a valuable limited resource. The existence of the Kuiper license acted as a deterrent to other potential applicants, and this is not in the public interest. If FCC allows the extension, they encourage all other spectrum squatters.

[deltaV]

At an absolute minimum, they should have contracted for a whole lot more than three F9s in December 2023

Are we sure that SpaceX was willing to sell Amazon "a whole lot more than three" launches for a reasonable price? The Starship delays presumably mean there are more Starlink Falcon launches than anticipated and that may naturally have produced a short-term Falcon supply shortage. SpaceX could probably increase Falcon capacity but this would take time and they may be reluctant to make this investment given the fact that they want to discontinue Falcon. SpaceX could bump Starlink to launch Kuiper but they may be reluctant to do so in order to launch a Starlink competitor (among other things the Starlink team would naturally feel they were being treated unfairly if that happened). I'm not saying that SpaceX definitely didn't give Amazon a good offer, just that we don't know if they did and shouldn't assume they did.

In any event SpaceX appears to have a monopoly on launch suitable for large constellations in the very short term and I think the FAA should avoid forcing anyone to buy from any (temporary or permanent) monopoly since monopolies are not in the public interest.

[DanClemmensen]

At an absolute minimum, they should have contracted for a whole lot more than three F9s in December 2023

Are we sure that SpaceX was willing to sell Amazon "a whole lot more than three" launches for a reasonable price? The Starship delays presumably mean there are more Starlink Falcon launches than anticipated and that may naturally have produced a short-term Falcon supply shortage. SpaceX could probably increase Falcon capacity but this would take time and they may be reluctant to make this investment given the fact that they want to discontinue Falcon. SpaceX could bump Starlink to launch Kuiper but they may be reluctant to do so in order to launch a Starlink competitor (among other things the Starlink team would naturally feel they were being treated unfairly if that happened). I'm not saying that SpaceX definitely didn't give Amazon a good offer, just that we don't know if they did and shouldn't assume they did.

In any event SpaceX appears to have a monopoly on launch suitable for large constellations in the very short term and I think the FAA should avoid forcing anyone to buy from any (temporary or permanent) monopoly since monopolies are not in the public interest.

Falcon 9 does not appear to be production-limited, but your point remains valid because it is still launch-rate-limited. Amazon would need to pay the full price-list price and probably pay more because they want to bump the queue since they did not make the order in a timely manner.

FAA has nothing to do with it. I assume you mean FCC. FCC's charter is orderly use of the airwaves, not LV business practices.

[gbl]

https://spaceflightnow.com/2024/02/12/cleanest-first-flight-ula-president-reflects-on-inaugural-vulcan-launch-and-future-of-program/

Bruno said once Vulcan starts flying Kuiper missions, it will launch 45 satellites with each flight. Atlas 5 rockets will carry 27 satellites each. Bruno said for these mega-satellite constellations, it’s important to get as many satellites on orbit as quickly as possible.

I believe up to now we were only speculating how many Kuiper satellites could be launched on ULA rockets. Now we can better calculate what launch rate is needed to meet the FCC deployment deadlines.

[DanClemmensen]

https://spaceflightnow.com/2024/02/12/cleanest-first-flight-ula-president-reflects-on-inaugural-vulcan-launch-and-future-of-program/

Bruno said once Vulcan starts flying Kuiper missions, it will launch 45 satellites with each flight. Atlas 5 rockets will carry 27 satellites each. Bruno said for these mega-satellite constellations, it’s important to get as many satellites on orbit as quickly as possible.

I believe up to now we were only speculating how many Kuiper satellites could be launched on ULA rockets. Now we can better calculate what launch rate is needed to meet the FCC deployment deadlines.

So,
   8 x 25 = 216 on Atlas V, leaving 1618-216=1402 more before 2026-07-30.
   1402/45= 32.1 Vulcan Centaur.
ULA needs to launch just over 40 Kuiper Launches from SLC-41 in 29 months, plus two or three Starliners, plus two other Atlas V.

[ThatOldJanxSpirit]

https://spaceflightnow.com/2024/02/12/cleanest-first-flight-ula-president-reflects-on-inaugural-vulcan-launch-and-future-of-program/

Bruno said once Vulcan starts flying Kuiper missions, it will launch 45 satellites with each flight. Atlas 5 rockets will carry 27 satellites each. Bruno said for these mega-satellite constellations, it’s important to get as many satellites on orbit as quickly as possible.

I believe up to now we were only speculating how many Kuiper satellites could be launched on ULA rockets. Now we can better calculate what launch rate is needed to meet the FCC deployment deadlines.

So,
   8 x 25 = 216 on Atlas V, leaving 1618-216=1402 more before 2026-07-30.
   1402/45= 32.1 Vulcan Centaur.
ULA needs to launch just over 40 Kuiper Launches from SLC-41 in 29 months, plus two or three Starliners, plus two other Atlas V.

Plus nine additional NSSL flights and two Dream Chaser flights manifested from SLC-41 by end 2025, and additional flights no doubt in H1 2026.

[Zed_Noir]

<snip>
So,
   8 x 25 = 216 on Atlas V, leaving 1618-216=1402 more before 2026-07-30.
   1402/45= 32.1 Vulcan Centaur.
ULA needs to launch just over 40 Kuiper Launches from SLC-41 in 29 months, plus two or three Starliners, plus two other Atlas V.

Plus nine additional NSSL flights and two Dream Chaser flights manifested from SLC-41 by end 2025, and additional flights no doubt in H1 2026.

Also LSP launches unless ULA don't bid on them.

There are the 3 Falcon 9 plus the 2 Ariane 62 and 16 Ariane 64+ launches booked by Amazon for deploying KuiperSats. With maybe x20 with the Falcon 9, maybe x15 with the Ariane 62 and x40 with the Ariane 64+ with the P140 solid boosters. Likely only a handful of Ariane 64+ launches will be before the regulatory deadline. The Falcon 9 and Ariane 62 launches are supposedly for 2025.

[StraumliBlight]

On March 8th, Kuiper requested to reduce its constellation by 4 satellites to 3,232:

 • 782 satellites at 590 km and 33°, with 1 satellite per plane
 • 2 satellites at 590 km and 30°, in 1 plane
 • 1292 satellites at 610 km and 42°, with 1 satellite per plane
 • 1156 satellites at 630 km and 51.9°, with 4 satellites per plane

[zubenelgenubi]

Posting this Amazon press release here for completeness:
Amazon's Project Kuiper completes successful tests of optical mesh network in low Earth orbit
December 14, 2023

After demonstrating 100 Gbps optical links between its prototype satellites, Project Kuiper will include laser links on every satellite in its constellation to form a mesh network in space.

Since the successful launch and deployment of two prototype satellites in October 2023, Project Kuiper has been conducting extensive testing of its end-to-end communications payload and network. In November, we confirmed we had validated all priority systems and subsystems within 30 days of launch, and we shared a first look at demos over the Kuiper network. One critical system that was part of that testing—but has remained confidential until now—was our optical inter-satellite link (OISL) capabilities.
Animation of Project Kuiper’s satellite constellation with optical links between satellites.

[Caption: Animation of Project Kuiper’s satellite constellation.]

Our prototype satellites, KuiperSat-1 and KuiperSat-2, include advanced optical communications payloads. We have completed multiple successful demonstrations of that next-generation technology, maintaining 100 gigabits per second (Gbps) links over a distance of nearly 621 miles (1,000 kilometers) for the entire test window. These tests validated the final component of Project Kuiper’s advanced communications architecture, and the results ensure that OISLs will be operational on our first production satellites, slated for launch in the first half of 2024.

“With optical inter-satellite links across our satellite constellation, Project Kuiper will effectively operate as a mesh network in space,” said Rajeev Badyal, Project Kuiper's vice president of technology. “This system is designed fully in-house to optimize for speed, cost, and reliability, and the entire architecture has worked flawlessly from the very start. These immediate results are only possible because we approached our OISL architecture as one part of a fully integrated system design, and it’s a testament to this team’s willingness to invent on behalf of customers. We’re excited to be able to support these next-generation OISL capabilities on every Kuiper satellite from day one.”

Operating a next-generation orbital laser mesh network

OISLs use infrared lasers to send data between spacecraft as they orbit the planet. Instead of being limited to sending data between an individual satellite and antennas on the ground, OISLs allow satellites to send data directly to other satellites in a constellation. We are equipping every Project Kuiper satellite with multiple optical terminals to connect many satellites at a time, establishing high-speed laser cross-links that form a secure, resilient mesh network in space. These capabilities increase throughput and reduce latency across our constellation, and provide more flexibility to connect Kuiper customers across land, sea, air, and space.

Another benefit of OISLs is the speed at which you can move data around the world. Lighttravels faster in space than it does through glass, which means that Kuiper’s orbital laser mesh network can move data approximately 30% faster than if it traveled the equivalent distance via terrestrial fiber optic cables. And because Project Kuiper leverages AWS service and infrastructure to route data traffic, we can reduce latency even further across our network.

Advancing the state of the art in OISL technology

OISLs represent a long-standing challenge for low Earth orbit (LEO) satellite broadband connectivity. To establish and maintain laser links, you need to minimize the spread of light to ensure a strong signal; you need to establish contact across distances of up to 1,616 miles (2,600 km); you need to maintain that connection between spacecraft moving at speeds of up to 15,534 miles per hour (25,000 km per hour); and you need to accomplish all of that while compensating for satellite and flight dynamics. Project Kuiper designed a state-of-the-art optics and control system capable of addressing these challenges.

[Caption: An image of an Amazon employee working on Project Kuiper prototype mission.

Project Kuiper tested OISL technology extensively in the lab before testing in space.]

Although this system performed well during extensive testing in a lab environment, our Protoflight mission provided an opportunity to test the end-to-end architecture in space. In early November, we established the first successful optical links between our two prototype satellites, allowing us to send and receive data at speeds of up to 100 Gbps for entire test windows of an hour or more. These tests demonstrated our ability to establish a single bi-directional link between two satellites, and initial data indicates that our design will be able to maintain cross-links between multiple satellites at once—the critical feature of a next-generation mesh network in space.

Delivering resilient, secure connectivity for enterprise and public sector customers

Amazon’s OISL capabilities will allow Project Kuiper to move and land data anywhere via its mesh network in space, bringing secure, resilient connectivity to a wide range of enterprise and public sector customers. This is especially important for those operating in regions without a nearby ground station—for example, a cruise liner in the middle of the ocean or an aircraft making a transatlantic flight—allowing them to securely uplink data from effectively any location on Earth, transmit it through space via laser communications, and downlink it to their destination of choice.

“Amazon’s optical mesh network will provide multiple paths to route data through space, creating resiliency and redundancy for customers who need to securely transport information around the world,” said Ricky Freeman, vice president of Kuiper Government Solutions, a dedicated division within Project Kuiper focused on serving public sector customers. “This is especially important for those looking to avoid communications architectures that can be intercepted or jammed, and we look to forward to making these capabilities available to public sector customers looking to move and land data from remote locations to their desired destination.”

Project Kuiper is already supporting public sector research and development in optical communications technology, and we look forward to bringing our advanced communications capabilities to additional enterprise and government customers as we begin to deploy our OISL-equipped production satellites.

Project Kuiper is beginning satellite manufacturing ahead of a full-scale deployment that starts in the first half of 2024, and we expect to have enough satellites deployed to begin early customer pilots in the second half of 2024.

To learn more about the Protoflight mission and next steps for the program, check out our latest mission updates.

[Asteroza]

I forget, are these OISL's from Mynaric or homemade?

[tssp_art]

Update on Kuiper in Geekwire Article - Kuiper satellite network gets a big lift from Amazon’s CEO, but timeline is a bit hazy

TLDR: After touting the successes of the first two prototype sats, Amazon/Kuiper announced that they would start sending up production satellites in the first half of this year; now they are saying some time later this year. They also said their broadband service would be in beta testing with selected customers by the end of this year; now they are saying they "expect to have the service up in the next year or so". They're behind schedule and falling further behind. There is a diminishing likelihood of meeting the FCC mandated schedule. Read the bolded paragraphs

In his annual letter to shareholders, Amazon CEO Andy Jassy says the company’s Project Kuiper satellite venture will be “a very large revenue opportunity” in the future — but he’s hedging his bets as to exactly when that future will be.

Eventually, Project Kuiper aims to provide satellite broadband service to hundreds of millions of people around the world who are currently underserved when it comes to connectivity. Such a service would compete with SpaceX’s Starlink satellite network, which already has more than 2.6 million customers.

Amazon is investing more than $10 billion to get Kuiper off the ground. The plan calls for sending 3,232 satellites (down slightly from the originally planned 3,236) into low Earth orbit by 2029. Under the terms of the Federal Communications Commission’s license, half of that total would have to be deployed by mid-2026.

When Project Kuiper’s first two prototype satellites were launched last October for testing, Amazon said that its first production-grade satellites were on track for launch in the first half of 2024, and that it expected broadband service to be in beta testing with selected customers by the end of the year.

Today, Jassy put a slightly different spin on that schedule. “We’re on track to launch our first production satellites in 2024,” he wrote in his letter. “We’ve still got a long way to go, but are encouraged by our progress.”

Jassy amplified on those remarks in an interview with CNBC. "The first big production pieces will be the second half of ’24, and we expect to have the service up in the next year or so," he said.

[snip]

Rajeev Badyal, vice president of technology at Project Kuiper, told GeekWire last November that his team had already identified the launch vehicles for the first production-grade satellites. "In the near future, we’ll start disclosing our launch plans — who we’re launching with, which rockets," he said. "We have a couple of options to start off with, but we’ve narrowed it for our first year, what we want to do. And we’ll share that information in the days to come."

[ThatOldJanxSpirit]

Since the October test sat launch we’ve seen Kuiper production move from late 2023 to H2 2024. It seems unlikely that we will see Kuipers flying on Atlas before Q3 and more likely Q4. So, apart from possibly a few Atlas flights late on 2024 is shaping up to be a bit of a washout for Kuiper.

Things don’t look to be much better for 2025. With Tory confirming that the Q4 Dreamchaser flight will be Cert 2 and USSF-106 now scraping in at the end of 2024 (possibly 2025, Wow!) pushing two or three more flights into the enormous bow wave of DoD launches.
Add to this mix the Atlas V Kuiper flights that won’t now be flown in 2024 and it seems unlikely that many Vulcan Kuiper missions will fly in 25 and indeed H1 26.
I don’t hold out much hope for Ariane 6 which is also progressing slowly and facing a backlog of government missions.

So it’s all up to Falcon and New Glenn now. More delays seem inevitable. My sweepstake call is only 800 Kuiper Sats operational by July 2026.

[DanClemmensen]

Since the October test sat launch we’ve seen Kuiper production move from late 2023 to H2 2024. It seems unlikely that we will see Kuipers flying on Atlas before Q3 and more likely Q4. So, apart from possibly a few Atlas flights late on 2024 is shaping up to be a bit of a washout for Kuiper.

Things don’t look to be much better for 2025. With Tory confirming that the Q4 Dreamchaser flight will be Cert 2 and USSF-106 now scraping in at the end of 2024 (possibly 2025, Wow!) pushing two or three more flights into the enormous bow wave of DoD launches.
Add to this mix the Atlas V Kuiper flights that won’t now be flown in 2024 and it seems unlikely that many Vulcan Kuiper missions will fly in 25 and indeed H1 26.
I don’t hold out much hope for Ariane 6 which is also progressing slowly and facing a backlog of government missions.

So it’s all up to Falcon and New Glenn now. More delays seem inevitable. My sweepstake call is only 800 Kuiper Sats operational by July 2026.

They also reserved eight Atlas V LVs, which are presumably ready to go as needed starting now. The fact that they are not using them points to issues with the satellites or satellite production rather than launcher availability. Did they discover a problem with the 2-satellite test flight? That's why you do test flights, after all. Eight Atlas V plus the 3 Falcon 9 are still nowhere near enough to launch 1616 satellites by July 2026. I don't know how far in advance you must reserve Falcon 9s, but maybe it's time to do that. SpaceX accommodated OneWeb in less than 9 months, but that was for their first of only 3 launches, and was a response to the unanticipated sudden loss of Soyuz launches. SpaceX has no such reason to be extra-accommodating for Kuiper, whose problems appear to be self-inflicted.

[JayWee]

I guess hiring people fired for being slow wasn't that bright idea...

[TrevorMonty]

They also reserved eight Atlas V LVs, which are presumably ready to go as needed starting now. The fact that they are not using them points to issues with the satellites or satellite production rather than launcher availability. Did they discover a problem with the 2-satellite test flight? That's why you do test flights, after all. Eight Atlas V plus the 3 Falcon 9 are still nowhere near enough to launch 1616 satellites by July 2026. I don't know how far in advance you must reserve Falcon 9s, but maybe it's time to do that. SpaceX accommodated OneWeb in less than 9 months, but that was for their first of only 3 launches, and was a response to the unanticipated sudden loss of Soyuz launches. SpaceX has no such reason to be extra-accommodating for Kuiper, whose problems appear to be self-inflicted.

Given satellite production seems to be limiting factor maybe now is not good time to book more LVs.

[DanClemmensen]

They also reserved eight Atlas V LVs, which are presumably ready to go as needed starting now. The fact that they are not using them points to issues with the satellites or satellite production rather than launcher availability. Did they discover a problem with the 2-satellite test flight? That's why you do test flights, after all. Eight Atlas V plus the 3 Falcon 9 are still nowhere near enough to launch 1616 satellites by July 2026. I don't know how far in advance you must reserve Falcon 9s, but maybe it's time to do that. SpaceX accommodated OneWeb in less than 9 months, but that was for their first of only 3 launches, and was a response to the unanticipated sudden loss of Soyuz launches. SpaceX has no such reason to be extra-accommodating for Kuiper, whose problems appear to be self-inflicted.

Given satellite production seems to be limiting factor maybe now is not good time to book more LVs.

This is equivalent to saying that Kuiper will not make the July 2026 deadline to launch 1616 satellites. That in turn means they must plan to get a waiver from the FCC. But I suspect that just about everybody in the industry, not just SpaceX, will oppose that.

[Kiwi53]

They also reserved eight Atlas V LVs, which are presumably ready to go as needed starting now. The fact that they are not using them points to issues with the satellites or satellite production rather than launcher availability.
{snip}

Given satellite production seems to be limiting factor maybe now is not good time to book more LVs.

This is equivalent to saying that Kuiper will not make the July 2026 deadline to launch 1616 satellites. That in turn means they must plan to get a waiver from the FCC. But I suspect that just about everybody in the industry, not just SpaceX, will oppose that.

How many Kuiper satellites would represent a Minimum Viable Product?
That might be enough so that at least one satellite is visible from any place in the USA (Hawaii & Alaska excepted). If they haven't even got to that by the July 2026 deadline, ISTM it would make their application to the FCC for an extension even more problematical

[RedLineTrain]

This is equivalent to saying that Kuiper will not make the July 2026 deadline to launch 1616 satellites. That in turn means they must plan to get a waiver from the FCC. But I suspect that just about everybody in the industry, not just SpaceX, will oppose that.

I wouldn't be so sure that many others will oppose a short waiver.  SpaceX probably will not, because its operations could be impacted the most in the future with any negative precedent.

[DanClemmensen]

They also reserved eight Atlas V LVs, which are presumably ready to go as needed starting now. The fact that they are not using them points to issues with the satellites or satellite production rather than launcher availability.
{snip}

Given satellite production seems to be limiting factor maybe now is not good time to book more LVs.

This is equivalent to saying that Kuiper will not make the July 2026 deadline to launch 1616 satellites. That in turn means they must plan to get a waiver from the FCC. But I suspect that just about everybody in the industry, not just SpaceX, will oppose that.

How many Kuiper satellites would represent a Minimum Viable Product?
That might be enough so that at least one satellite is visible from any place in the USA (Hawaii & Alaska excepted). If they haven't even got to that by the July 2026 deadline, ISTM it would make their application to the FCC for an extension even more problematical

If your LEO constellation can continuously serve any one point at a certain latitude, it can serve every point at that latitude (assuming a ground station or ISL) so no advantage to skipping Hawaii. You can skip Alaska but only by skipping Canada, Nordic countries, Russia,and Antarctica.