Amazon Project Kuiper Broadband Constellation

[meekGee]

TrevorMonty and meekGee: please use the "preview" button before posting and fix your quotes.

Hmm made the BB codes in the OP balance out, though I don't know if they're correct...

[Tywin]

Everyday, more powerful Amazon, and with Kuiper they will be a great competence...

[meekGee]

I get the feeling you're not a Costco card holder...

Though by your own graphs  Amazon is operating at very thin margins - $540B turnover, and $30B profit.  That's 5.5%.  (And that's after the rebound)

Meanwhile the second slide says MS is closing in, probably a lot because Amazon fumbled with AI.

AI is very capital intensive.  The compute farms needed are not the same as the servers AWS has.

And Kuiper is capital intensive too  especially without a launch vehicle that can complete with Starship.

In short: Amazon is behind on the two cutting edge fronts, and might have to choose between the two.

I think if it came down to it, they'll choose AI over a satellite constellation.

Remember what happened to Amazon phone, which could have been even more strategic. It was late, it didn't get traction, we got canceled.

[Tywin]

IA is a buble at the moment, and pure smoke with a lot of PR...

Where is my car that drive alone, and safe?

Where is my robot that make laundry, clean cars, cooking, etc...

And Amazon have here own IA project, like others...

No I don't have Costco card, but I check the number of amazon, and in Europe, they will grow a lot in the e-commerce service...

Plus, Prime ads, Twich, Kuipersat, Amazon Music, etc...and the golden chicken AWS...

[Tywin]

I get the feeling you're not a Costco card holder...

Though by your own graphs  Amazon is operating at very thin margins - $540B turnover, and $30B profit.  That's 5.5%.  (And that's after the rebound)

Meanwhile the second slide says MS is closing in, probably a lot because Amazon fumbled with AI.

AI is very capital intensive.  The compute farms needed are not the same as the servers AWS has.

And Kuiper is capital intensive too  especially without a launch vehicle that can complete with Starship.

In short: Amazon is behind on the two cutting edge fronts, and might have to choose between the two.

I think if it came down to it, they'll choose AI over a satellite constellation.

Remember what happened to Amazon phone, which could have been even more strategic. It was late, it didn't get traction, we got canceled.

Nothing comprare with Microsoft phone, a complete disaster even more after buy Nokia phone...

[TrevorMonty]

Amazon tends to pour alot of their profit back into growing company with R&D being one of the biggest investments. Is that $30B before or after reinvestment into the company has been deducted?

While AWS maybe losing market share doesn't mean they are losing customers and business is shrinking. Microsoft market gain in cloud server market is likely to be on new business.

The most important thing is that business is growing every year not how well competition is doing.

[meekGee]

I get the feeling you're not a Costco card holder...

Though by your own graphs  Amazon is operating at very thin margins - $540B turnover, and $30B profit.  That's 5.5%.  (And that's after the rebound)

Meanwhile the second slide says MS is closing in, probably a lot because Amazon fumbled with AI.

AI is very capital intensive.  The compute farms needed are not the same as the servers AWS has.

And Kuiper is capital intensive too  especially without a launch vehicle that can complete with Starship.

In short: Amazon is behind on the two cutting edge fronts, and might have to choose between the two.

I think if it came down to it, they'll choose AI over a satellite constellation.

Remember what happened to Amazon phone, which could have been even more strategic. It was late, it didn't get traction, we got canceled.

Nothing comprare with Microsoft phone, a complete disaster even more after buy Nokia phone...

Exactly! Further illustrating my point.

"Infinite" financial resources backing up a project (whether by Amazon or Microsoft or anyone) can't help if the project can't find its footing and help itself.  Even those infibite resources become finite at some point.

Kuiper is starting off many years late, and with a much lower and more expensive set of launchers. It'll be a hard sell.

As for AI, while self driving cars remain (and probably will remain) a far off dream, the big money is in natural language processing and "understanding".  Basically all those tasks that human would say require "zero intelligence", and yet computer algorithms were unable to perform.

It's like the advances of image classification, but applied to human conversation.  That's really big.

I know AWS is absolutely scrambling to catch up and hire talent in that area. They absolutely did not see it coming, and they are losing clients to MS and Google cloud because both of those are offering AI services to hosted platforms.

I am not aware of AWS losing customers to anyone because they don't have a satellite constellation.

If I was an Amazon board member, looking to sink 10s of Billions into either telecom or AI, the choice would be easy.

[meekGee]

Amazon tends to pour alot of their profit back into growing company with R&D being one of the biggest investments. Is that $30B before or after reinvestment into the company has been deducted?

While AWS maybe losing market share doesn't mean they are losing customers and business is shrinking. Microsoft market gain in cloud server market is likely to be on new business.

The most important thing is that business is growing every year not how well competition is doing.

When your nearest competitor is some 2/3 your size, comparative trends matter.

Amazon'a willingness to spend money on Kuiper (which is the on-topic connection) will be influenced by:
- How expensive it is
- What are its advantages to AWS
- What are its odds of making money as a business unit
- How much spare money does Amazon have
- What other projects are demanding investment money

So:
- Very
- Not clear really. Other cloud services aren't going for it, and Starlink is neutral and exists.
- Not very good, in light of competitive disadvantage to Starlink
- Plenty but not infinite
- AI is the most obvious example, also in light of Amazon's investment in customer interaction (e.g. Alexa), and the fact that its lack (in comparison with other cloud providers) is hurting AWS

Much to consider here wrt to Kuiper's guaranteed income.

[RedLineTrain]

Amazon tends to pour alot of their profit back into growing company with R&D being one of the biggest investments. Is that $30B before or after reinvestment into the company has been deducted?

While AWS maybe losing market share doesn't mean they are losing customers and business is shrinking. Microsoft market gain in cloud server market is likely to be on new business.

The most important thing is that business is growing every year not how well competition is doing.

Important to take Amazon's profit and loss statements with a grain of salt.  In the extreme, all spending, including for the product (a product discount), can be said to be an investment in the business.  So gross profit, net profit, R&D spending, etc. are only marginally useful figures of merit.  The "property, plant, and equipment" cash flow line item is probably the most salient as it relates to AI and Kuiper spending.

In the last twelve months, Microsoft spent $35 billion on property, plant, and equipment and is trending toward $40-$50 billion spending a year.  The vast majority of this spending probably is on data centers, and a majority of that spending seems to be on AI.

In the last twelve months, Amazon has spent $53 billion on property, plant, and equipment.  This is down a bit from prior years, in which the company was spending heavily on support for same-day and one-day shipping.  Only a portion of this is devoted to AWS.  From 2011 through 2022, that was roughly $10 billion a year.  But it wouldn't surprise me to learn that 2023 was higher, as it is on the back foot in competition with Microsoft on AI.

Kuiper is probably a roughly $5 billion a year spending item.  Quite substantial, if you compare to AWS spending.  I can see heightened scrutiny of it, and wouldn't be surprised to see them pull the plug on it, even if that's not the most likely outcome.  Microsoft's business generates more cash than does Amazon's.  So in extreme circumstances, Microsoft could outspend Amazon, leading Kuiper to fall by the wayside.  I do feel like Kuiper is a "nice to have" rather than "core business need."

[TrevorMonty]

Microsoft us also on playing space domain. Via Starlink and other things see article.
 
https://www.geekwire.com/2022/microsoft-azure-access-cloud-spacex-starlink-satellites/

Will interesting to see how their space side of business grows.

[meekGee]

Microsoft us also on playing space domain. Via Starlink and other things see article.
 
https://www.geekwire.com/2022/microsoft-azure-access-cloud-spacex-starlink-satellites/

Will interesting to see how their space side of business grows.

Yup so they chose to use Starlink rather than develop their own constellation.

The only reason I can see for Amazon to develop Kuiper is to give NG something to do - and that's not supposed to happen because theoretically they're different companies following independent agendas.

In practice, this will only carry so much weight. At some point Amazon will be forced to do the math.

Note that in the article it says MS is offering the service now (using Starlink), but Amazon is waiting for Kuiper. Which is waiting for launch vehicles.

That's an example of Amazon hurting itself because of allegiance to BO.

[DanClemmensen]

Microsoft us also on playing space domain. Via Starlink and other things see article.
 
https://www.geekwire.com/2022/microsoft-azure-access-cloud-spacex-starlink-satellites/

Will interesting to see how their space side of business grows.

Yup so they chose to use Starlink rather than develop their own constellation.

The only reason I can see for Amazon to develop Kuiper is to give NG something to do - and that's not supposed to happen because theoretically they're different companies following independent agendas.

In practice, this will only carry so much weight. At some point Amazon will be forced to do the math.

Note that in the article it says MS is offering the service now (using Starlink), but Amazon is waiting for Kuiper. Which is waiting for launch vehicles.

That's an example of Amazon hurting itself because of allegiance to BO.

I can think of one reason that Amazon might think they need their own network: real-time control of delivery drones. Dependency on a third party network might affect the regulatory situation. This is not technical issue: the third party might provide better service than Kuiper. The issue is who would guarantee the required service levels.

You might think this is a niche and it is as far as Starlink in concerned. But if Starlink is unwilling to provide the needed contractual service commitments, then Amazon might need Kuiper for this, and if they must implement Kuiper, then they would use it for all of their own traffic (delivery trucks, etc.) and also sell it to other users.

I have zero reason to believe that this is the reason Amazon is building Kuiper. It's just idle speculation, as I desperately search for a justification for this insane project.

[Asteroza]

Microsoft us also on playing space domain. Via Starlink and other things see article.
 
https://www.geekwire.com/2022/microsoft-azure-access-cloud-spacex-starlink-satellites/

Will interesting to see how their space side of business grows.

Yup so they chose to use Starlink rather than develop their own constellation.

The only reason I can see for Amazon to develop Kuiper is to give NG something to do - and that's not supposed to happen because theoretically they're different companies following independent agendas.

In practice, this will only carry so much weight. At some point Amazon will be forced to do the math.

Note that in the article it says MS is offering the service now (using Starlink), but Amazon is waiting for Kuiper. Which is waiting for launch vehicles.

That's an example of Amazon hurting itself because of allegiance to BO.

I can think of one reason that Amazon might think they need their own network: real-time control of delivery drones. Dependency on a third party network might affect the regulatory situation. This is not technical issue: the third party might provide better service than Kuiper. The issue is who would guarantee the required service levels.

You might think this is a niche and it is as far as Starlink in concerned. But if Starlink is unwilling to provide the needed contractual service commitments, then Amazon might need Kuiper for this, and if they must implement Kuiper, then they would use it for all of their own traffic (delivery trucks, etc.) and also sell it to other users.

I have zero reason to believe that this is the reason Amazon is building Kuiper. It's just idle speculation, as I desperately search for a justification for this insane project.

There is a sort of market for this currently actually, in different forms. One is remote "safety" drivers for various UGV's in various forms. UGV's could be fairly autonomous, but if they get into something too complex, signal the need for human intervention. There have been various forms of this already. Rio Tinto was trialing remote driven mine trucks. Tractors driven remotely for understaffed farms (John Deer is going to add Starlink to their tractors soon)(Komatsu is well known for having remote vehicle informatics on their construction equipment). Some autonomous truck companies operate autonomous on highways but have remote human drivers for city streets. There are psuedo chauffeur services now where a rental car is remotely driven to the customer (who then drives manually themselves), and retrieved remotely afterwards (this is an interesting intermediary step, next being remote driven taxis with live occupants, then remote supervised autonomous taxis, then robotaxis with only remote intervention)

Amazon is currently heavily instrumenting their company owned electric delivery vans, ostensibly to collect training data for autonomous driving purposes. A future scenario where you can schedule curbside delivery by autonomous cargo vans, where you open a side door to pick up your package is easily understood and can be almost done right now. Avoids the last 100 feet problem of getting a package from the van to your front door, though Amazon is doing some work with humanoid type robot companies to cover the van-to-door issue. Cutting back redelivery by using electric vans that can loiter or park in neighborhood areas is an interesting model.

[TrevorMonty]

I have zero reason to believe that this is the reason Amazon is building Kuiper. It's just idle speculation, as I desperately search for a justification for this insane project.

So a LEO broadband constellation is an insane project. Maybe you should tell Elon that.

[Zed_Noir]

I have zero reason to believe that this is the reason Amazon is building Kuiper. It's just idle speculation, as I desperately search for a justification for this insane project.

So a LEO broadband constellation is an insane project. Maybe you should tell Elon that.

Only insane if you are way behind the second mover (OneWeb). Not so insane if you are the first mover and has gobble up most of the market share. To the point of deploying competitors' LEO constellations without too much worries.

I prefer the simplest explanations for Project Kuiper. Some bald guy has some bad presumptions on how easy it will be and isn't able to quit once Kuiper got started.

[meekGee]

I have zero reason to believe that this is the reason Amazon is building Kuiper. It's just idle speculation, as I desperately search for a justification for this insane project.

So a LEO broadband constellation is an insane project. Maybe you should tell Elon that.

The only unclear part is why Amazon needs its own, not why it exists at all.  A LEO constellation clearly makes sense.

[DanClemmensen]

I have zero reason to believe that this is the reason Amazon is building Kuiper. It's just idle speculation, as I desperately search for a justification for this insane project.

So a LEO broadband constellation is an insane project. Maybe you should tell Elon that.

Kuiper is insane. Starlink is not insane. The difference is that Starlink is already in place with no effective competition. It got there by being one half of SpaceX' fully integrated business plan that included Falcon 9. OneWeb had a chance, but they lost their initial cheap access to space (Soyuz) and they had no launch business to fund their Constellation.

[TrevorMonty]

I have zero reason to believe that this is the reason Amazon is building Kuiper. It's just idle speculation, as I desperately search for a justification for this insane project.

So a LEO broadband constellation is an insane project. Maybe you should tell Elon that.

Kuiper is insane. Starlink is not insane. The difference is that Starlink is already in place with no effective competition. It got there by being one half of SpaceX' fully integrated business plan that included Falcon 9. OneWeb had a chance, but they lost their initial cheap access to space (Soyuz) and they had no launch business to fund their Constellation.

Thank goodness entrepreneurs don't have your defeatist attitude, other wise all our product and services suppliers would have monopoly.
Customers switch suppliers regularly for varies reasons g poor service, price,  features. I've switched  broadband providers of broadband, power,  banks and insurance companies on a few occasions for all of those reasons.

One particular automotive manufacturer startup 20 odd years ago took on established markets leader. They now have significant piece of the market.

[DanClemmensen]

I have zero reason to believe that this is the reason Amazon is building Kuiper. It's just idle speculation, as I desperately search for a justification for this insane project.

So a LEO broadband constellation is an insane project. Maybe you should tell Elon that.

Kuiper is insane. Starlink is not insane. The difference is that Starlink is already in place with no effective competition. It got there by being one half of SpaceX' fully integrated business plan that included Falcon 9. OneWeb had a chance, but they lost their initial cheap access to space (Soyuz) and they had no launch business to fund their Constellation.

Thank goodness entrepreneurs don't have your defeatist attitude, other wise all our product and services suppliers would have monopoly.
Customers switch suppliers regularly for varies reasons g poor service, price,  features. I've switched  broadband providers of broadband, power,  banks and insurance companies on a few occasions for all of those reasons.

One particular automotive manufacturer startup 20 odd years ago took on established markets leader. They now have significant piece of the market.

I'm not negative about all new companies or new initiatives. I am negative about Kuiper. The company I started used new technology to successfully enter a market that was dominated by old-line suppliers, and we won nearly all new bids. It can be done, but you need some sort of differentiator or a major advantage in price or functionality.

[meekGee]

I have zero reason to believe that this is the reason Amazon is building Kuiper. It's just idle speculation, as I desperately search for a justification for this insane project.

So a LEO broadband constellation is an insane project. Maybe you should tell Elon that.

Kuiper is insane. Starlink is not insane. The difference is that Starlink is already in place with no effective competition. It got there by being one half of SpaceX' fully integrated business plan that included Falcon 9. OneWeb had a chance, but they lost their initial cheap access to space (Soyuz) and they had no launch business to fund their Constellation.

Thank goodness entrepreneurs don't have your defeatist attitude, other wise all our product and services suppliers would have monopoly.
Customers switch suppliers regularly for varies reasons g poor service, price,  features. I've switched  broadband providers of broadband, power,  banks and insurance companies on a few occasions for all of those reasons.

One particular automotive manufacturer startup 20 odd years ago took on established markets leader. They now have significant piece of the market.

That particular automaker entered the market using completely new technologies and disrupted it.

What Kuiper is doing is akin to entering the automotive market today with a product line that relies on internal combustion engines.

People will gladly switch vendors, that's true, but only if the vendors are comparable or if the new vendor offers an advantage. How will Kuiper even be comparable to Starlink?