SpaceX vs. Everyone - ULA, NG, Boeing, Lockheed, etc.

[scaesare]

And...

Amazon is facing a schedule crunch to launch its Kuiper satellites. Its Federal Communications Commission license requires the company to have half the constellation, or more than 1,600 satellites, in orbit by July 2026, and the full constellation in orbit by July 2029.

And according to Amazon's Kuiper Site they were originally planning to launch two prototype sats in "early 2023", but an update at the top of the page says now targeting Fall '23. That starts in 3 weeks... anybody aware of if they are even close?

And if those are prototype sats, what are the chances of 1600 hundred birds flying in less than 3 years with needing to finalize the sat design, go into mass production, and the launch services available?

Time for another bet thread??

 

[Cosmacelf]

And then The obvious “While the value of the Atlas launch contract is redacted in the suit, it argues that SpaceX’s list price for Falcon 9 launches was significantly less.”

In general I don’t like lawsuits against companies for management mistakes, but this one does reek of double dealing and Amazon shareholders were indeed screwed over. Actually the whole Kuiper idea is a massive bad idea for Amazon.

 

[JB47394]

anybody aware of if they are even close?

From spaceflightnow.com

NET September 26 Atlas 5 • Project Kuiper
Launch time: TBD
Launch site: SLC-41, Cape Canaveral Space Force Station, Florida
The first two demonstration satellites for Amazon’s Project Kuiper broadband constellation will launch on an Atlas 501 rocket. These satellites were originally scheduled to fly on the first Vulcan rocket.
Updated: August 20

So they're still at "No Earlier Than".

And if those are prototype sats, what are the chances of 1600 hundred birds flying in less than 3 years with needing to finalize the sat design, go into mass production, and the launch services available?

A Kuiper satellite is supposed to mass 600-700kg, so we're talking 960-1120 tons. The Vulcan Centaur (VC6) can loft 27 tons to LEO, allowing 38-45 satellites per flight. For 1600 satellites, that's 36-42 flights. I don't know what determines the launch cadence for rockets like that, but the Atlas V was usually launching about once per month. That said, there have been times when a pair of rockets launched within 8 days. I'm going to assume one flight per month for Kuiper satellites, which makes it three to four years just for launches. That's assuming everything goes well.

Of course, if they can build and launch Vulcan Centaurs faster, and the Kuiper satellites are being cranked out to keep up, then they can bring that timeline in. I don't know how important it is to hit that date.

 

[Cosmacelf]

They won’t hit that FCC date of course. Options: a) reduce constellation size, b) scrap project, c) bribe the FCC to extend the timeframe. I’m going with c).

 

[Electroman]

Actually the whole Kuiper idea is a massive bad idea for Amazon.

I would sue for this. What is the benefit for Amazon the retailer on this space internet project ? it is just another meaningless ego trip for Bezos.

 

[hmcgregoraz]

I would sue for this. What is the benefit for Amazon the retailer on this space internet project ? it is just another meaningless ego trip for Bezos.

The supposed benefit is for interconnecting their datacenter/hosting/cloud business.

For example, Google owns about 10,000 route miles of undersea fiber optics.

Facebook/meta is also involved in a fair amount of undersea fiber.

Amazon, and Microsoft also own a fair amount of undersea fiber.

The idea with this project for Amazon is to get additional users of their cloud services, and provide lower latency interconnects for their datacenters.

The primary goals are not related to the retail side of Amazon, but I am sure they won't have a problem selling off excess capacity to consumers.

-Harry

 

[Cosmacelf]

The supposed benefit is for interconnecting their datacenter/hosting/cloud business.

For example, Google owns about 10,000 route miles of undersea fiber optics.

Facebook/meta is also involved in a fair amount of undersea fiber.

Amazon, and Microsoft also own a fair amount of undersea fiber.

The idea with this project for Amazon is to get additional users of their cloud services, and provide lower latency interconnects for their datacenters.

The primary goals are not related to the retail side of Amazon, but I am sure they won't have a problem selling off excess capacity to consumers.

-Harry

Sounds to me like a very expensive way to connect datacenters. Fiber has got to be way cheaper in comparison.

 

[hmcgregoraz]

Sounds to me like a very expensive way to connect datacenters. Fiber has got to be way cheaper in comparison.

Agreed, but the latency of LEO based laser interconnect is theoretically lower than the latency of fiber.

Some financial customers will pay significantly for lower latency. These are the types of Cloud customers Amazon dearly wants to win.

There is a microwave network between NYC and Chicago specifically for the latency advantage over fiber, also London and Frankfurt.

https://arstechnica.com/information-technology/2016/11/private-microwave-networks-financial-hft/

-Harry

 

[scaesare]

I took that as offering service, using the Starlink sats with some additional security/QoS.

Not exactly, it would leverage the Starlink technology, but it doesn't just offer secure communication, as the page says the current focus of Starshield is 3 areas: Earth Observation, Communications and Hosted Payloads. Note: a. Currently Starlink constellation doesn't do Earth Observation; b. the Hosted Payload could be anything, ultimately everything DoD wanted could be a hosted payload.

Whether this will be a service is not clear, I don't think giving control of the Starshield constellation to DoD is out of the question. It could be on-orbit delivery of satellites, and then train DoD personnel to use SpaceX built ground systems.


They built 4 prototype missile tracking satellites for the SDA constellation, I assume this is what you mean by "build to spec", but they told SDA they're not interested in doing this anymore.

"Build to spec" is not SpaceX's MO, they're more interested in coming up with their own spec and controlling everything end to end (i.e. vertical integration). Remember Elon's first step in design process is "Make the requirements less dumb", hard to do this in "build to spec".

Revisiting this since more is being posted about Starshield since they recently landed the DoD contract... including this Verge article that quotes an Air Force spokesman as saying:

"Starshield end-to-end service via the Starlink constellation, user terminals, ancillary equipment, network management and other related services.”

So perhaps this is service layered on top of the existing constellation... or the person mis-spoke...

 

[JB47394]

So perhaps this is service layered on top of the existing constellation...

The contract is only $70 million for one year. It has to be using Starlink. I suspect it's just a formalization of the Ukraine situation though they may also be fooling with it to see how it works out in the field for the US military. Perhaps letting Ukraine continue to use it is part of their "field testing".

 

[Cosmacelf]

Yeah, I suspect this is an interim solution. An actual purpose built Starshield would be larded up with all sorts of sensors and gizmos, so will be a much more long term and expensive contract.

 

[TunaBug]

Yeah, I suspect this is an interim solution. An actual purpose built Starshield would be larded up with all sorts of sensors and gizmos, so will be a much more long term and expensive contract.

Care to elaborate on what those sensors and gizmos might be?

For military use I hope the traffic is isolated onto their private satellites. But from a business & engineering standpoint, wouldn't it be great if implementing those private services was nothing more than a VLAN and tweaks to some QOS rules? Which is what I infer from "Starshield end-to-end service via the Starlink constellation..."

 

[miimura]

I would hope that DoD would use terminals with their own encryption (that is controlled by DoD) built in and not just rely on Starlink VLANs.

 

[Cosmacelf]

Care to elaborate on what those sensors and gizmos might be?

I have no idea, I just find it rare to have a military procurement that is simple and zero extra requirements.

 

[JB47394]

Care to elaborate on what those sensors and gizmos might be?

Cameras would be the obvious starting point. Stick a good commercial sensor ($100) onboard and you can resolve down to 8 meters from 500km up (well, plus optics). Get serious about your camera and you can do far better. But imagine a bunch of Starlink satellites imaging the same spot from multiple angles as they fly by. You'd get real time stereo (or better) viewing.

Another would be antennas to pick up raw signals that are then broadcast to ground stations for processing, but I know precious little about what a small satellite could do with that.

The real appeal of a Starlink satellite constellation is that it allows for 24 hour coverage of the entire planet. For anyone who has read Ringworld, this is something that I've been looking forward to since I was a teen. Unfortunately, now that I'm an adult I know full well how people will abuse the power, just as there's always someone ready to abuse any power.

 

[scaesare]

Revisiting this since more is being posted about Starshield since they recently landed the DoD contract... including this Verge article that quotes an Air Force spokesman as saying:

"Starshield end-to-end service via the Starlink constellation, user terminals, ancillary equipment, network management and other related services.”

So perhaps this is service layered on top of the existing constellation... or the person mis-spoke...

And more from SpaceX providing Starlink services to DoD under ‘unique terms and conditions’ :

“The task order for Starshield services is provided by the Starlink satellite constellation but is differentiated from the commercial Starlink service based on unique Department of Defense terms and conditions that are not found in commercial service contracts,” said the spokesperson.

So it looks like my earlier assumption this was a service layered on top of the existing constellation was correct, and it's not a seperate constellation as asserted:

They already are, their military constellation offering is called Starshield: SpaceX - Starshield

There's a unconfirmed rumor that they're offering DoD an 800 Starshield constellation for $2B. This would be a separate constellation from the Space Development Agency (SDA) constellation, as you can see the SDA constellation is already becoming a handout to old space military contractors.

 

[spacerfc]

And more from SpaceX providing Starlink services to DoD under ‘unique terms and conditions’ :

So it looks like my earlier assumption this was a service layered on top of the existing constellation was correct, and it's not a seperate constellation as asserted:

It's both, they're selling commercial Starlink service to military under the label Starshield, they're also planning or building a dedicated military constellation called Starshield. Or put it another way, Starshield is simply the brand name for their military products.

 

[scaesare]

It's both, they're selling commercial Starlink service to military under the label Starshield, they're also planning or building a dedicated military constellation called Starshield. Or put it another way, Starshield is simply the brand name for their military products.

Do you have any references to the dedicated constellation plans?

 

[spacerfc]

Do you have any references to the dedicated constellation plans?

Elon mentioned it on twitter: https://x.com/elonmusk/status/1700345943105638636

SpaceX is building Starshield for the US government, which is similar to, but much smaller than Starlink, as it will not have to handle millions of users. That system will be owned and controlled by the US government.

Also if you look at SpaceX - Starshield, it implied Starshield has separate satellites from Starlink, for example: "Starshield satellites are capable of integrating a wide variety of payloads, offering unique versatility to users."

And you get the same feel if you look at the Starshield jobs they're posting, it feels like a separate system with its own hardware stack, for example there're Starshield engineering positions for antenna, avionics, build/manufacturing, electrical, DSP, GNC, RF, etc.

There're some additional rumors online, but the above are the official sources.

 

[dhanson865]

I'd expect both layers to be up at the same time. The US military has a large unclassified network in addition to the classified network.

We already know the consumer grade setup can speak to more than one satellite at a time to handle switchovers and route data the fastest way possible.

I'm sure the military grade units will be able to handle communicating to an unclassified Starlink sat and a classified Starshield sat at the same time.

So I expect Starshield will route unclassified data over the existing Starlink network and classified over the new Starshield network. The terminal will be capable of connecting to either type of satellite and routing the data to the appropriate network. Use of both physical networks will be branded Starshield even though some traffic will be on the old Starlink network.