Hacker News new | past | comments | ask | show | jobs | submit login
Amazon investing $10B to compete against SpaceX in satellite broadband (arstechnica.com)
93 points by amaajemyfren 49 days ago | hide | past | favorite | 65 comments

Related question: What's the tech here that this can be accomplished (streaming data via satellite) without spectrum? Spectrum assets are some of the most valuable assets in the US (and are owned by Verizon et al). Satellite phone networks (LightSquared is the big one) have been stymied in the past by unavailability of spectrum and US Govt. concerns about interfering with GPS.

Is that not a problem in this case?

They're using frequencies in the 10.7 to 30.0 GHz range (GPS is much lower at 1.575 GHz) that are already allocated to satellite operations. However, they do have to share these frequencies with others. You can read all the frequency stipulations in the FCC authorization.


For example, operation at 27.5 to 28.35 GHz overlaps a 5G millimeter band in the US. 5G is primary and satellite uplink is secondary in that band, so if Kuiper interferes with a 5G installation, they can be ordered to mitigate the interference (possibly by stopping transmissions).

Lightsquared bought satellite spectrum close to gps, and then did some shady stuff to get terrestrial use approved. This was an attempt to arbitrage the difference between cheap satellite spectrum and expensive terrestrial spectrum. It should never have gotten as far as it did but businessmen and regulators didn't understand the science/technology.

Since this story came out a few days ago, does anyone know if Amazon has commented on whether or not they plan to launch these 'themselves' / with Blue Origin?

Right, seems like the first question people would be interested in hearing an answer to. I assume there is no answer yet. With a quick check, I was surprised to see that Blue Origin is developing a rocket which would be capable of launching these things (launch in 2021?)

But could they launch as cheaply or as often as SpaceX's Falcon 9, which is years ahead?

SpaceX isn't resting on its laurel either, with their R&D effort on Starship.

Maybe they don't need to launch as cheaply. SpaceX wasn't the first to the idea. There are other players and these companies didn't have their own rockets. The numbers must be good enough to work even with pre-SpaceX launch costs.

Elon said that the satellites are now cheaper to build than the cost to get them to orbit, and that they need starship to make it financially viable, iirc.

He says this during a keynote with the US Air Force about 4 months ago. He also says a ton of other interesting things as well.

Who says Amazon might not use SpaceX to bring the satellites in orbit?

Jeff. I think he would rather personally subsidize every New Glenn launch than letting Amazon launch their satellites with SpaceX.

What’s that based on? Is there bad blood there?

Jeff Bezos himself is owner of a rocket company called "Blue Origin". This company is developing a heavy-lift rocket awhich hasn't flown yet though. They already have some experience with booster return, as they have a small(er) vehicle serving as (future) science and tourism platform that can shoot the capsule with passengers and/or payload straight up over the Kármán line and have the booster return to the launch site.

Since they already had achieved this, Bezos tweeted "Welcome to the club" when Elon Musk's SpaceX had their first successful rocket landing. If that's appropriate is subject for debate, as the conditions experienced are not comparable.

Blue Origin also tried to patent landing a rocket booster on a boat while SpaceX was planning to exactly do that which lead to some bad blood.

I think Blue Origin is more focused towards possibly building a lunar lander. The launches I have see were with comparable tiny rockets, unless they launch thousands of them, they won't carry the satellites needed in time. Curious to get more informed opinions on this actually.

"New Shepard" is only for tourists or scientists running experiments in the capsule.

Any real payload would be launched with "New Glenn": https://en.wikipedia.org/wiki/New_Glenn

Thanks for sharing that, seems to still be out a bit - it took SpaceX quite a while to make Falcon 9 solid. So, if New Shephards first launch is next yearz it will probably take another 2-4 years until all glitches that appear are fixed. So I wouldn't be surprised if they send the first batches of satellites using rockets of other manufacturers.

The reality is, that in the current space market, if you don't have some way to produce launches for yourself, your gone have a bad time.

The GTO market, long the stable of many big rockets, is slowing and not gone pick up.

At the same time every government tries to prop up its domestic industry. One the few commercial launches, beating SpaceX on price is just not gone happen unless you simply fly with a massive lost, you need a couple of years ramp at least.

The other big thing it government launches, SpaceX and ULA are very established and its pretty unlikely that Blue will get deep into these anytime soon. SpaceX is gone continue to have most commercial Space Station buissness. ULA and SpaceX will likely again be selected for the military over the next couple years.

One of the reason SpaceX does Starlink is that it is an amazing driver for scale and re-usability for their rocket buissness. This year alone SpaceX without Starlink does only a few launches, not enough, and this is a year where most others did very few launches.

Blue and Bezos will have made the same calculation, you simply can't New Glenn sized rockets without that.

In every instance you said "gone" you should've said "going"

One of the more interesting use cases for SpaceX/Starlink and Amazon’s satellite internet projects is their use as a back haul (not sure if the correct term) provider for other ISPs as well.

One of the biggest roadblocks to starting a new ISP or expanding an existing ISPs service area is the big capital expenditure of wiring up areas. I would imagine that having the ability to rent capacity on these networks of satellites would drastically decrease the barrier to entry for new ISPs to form.

I can also see it being useful for mobile data coverage in remote areas. You could install a cell tower along a highway somewhere remote, and you’d just need a solar panel to power the tower, satellite signal to get internet coverage, and then the cell radio to broadcast that to phones.

Interesting tweet about interference and collisions between the two systems.


Link to FCC filing.


I would like to see the economics of this. There is a saving in not having to have 5g towers but if everyone in a small village has to spend $$$ on their own satellite dishes then the economics aren't going to make rural high speed internet via satellite so good.

Plus the uplink/downlink kit will be beyond mass consumer gear so the economies of scale won't be great.

Anyone seen the uplink kit for the SpaceX or British efforts?

Adopting broadband from Amazon or the other tech giants like google seems like a bad idea for society. They have too much power to begin with, and given allegations of how amazon may have abused seller data, I wouldn’t want them to have access to even more data. Not to mention issues around centralized control over information and the digital town square.

...launched with what rocket exactly? I'll believe them (and Blue Origin, supposedly) only after they put something up in space.

Blue Origin's timelines are actually realistic, so I wouldn't be surprised to see New Glenn (closer in payload to Falcon Heavy than Falcon 9) actually launch next year.

in what sense are they realistic?

I would be happy if they were on time, but it seems BO missed their milestones quite often.

Hitting objectives is way more important than hitting a date in BO's case. Even if SpaceX has a head start, Amazon has scale.

Blue Origin is not amazon.

When Amazon invests 10B it is. That is like saying Alphabet isn't Google.

Jeff Bezos invests privately, not Amazon (of course due to that relationship talent and resources will probably flow more easily between the companies).

It's like saying SpaceX isn't Tesla.

I’m wondering why it’s FCC which need to approve satellite communication services which orbits and provides services to the entire world?

I mean is the LEO is open for all countries to launch as many satellites as they want? What if China launches 10k satellites in the next year hypothetically?

Because they still need ground stations (and users!) in the US?

Are these supposed to be replacement of Satellite Phones? If I have a cheap Gbps connection or a cheaper 100 Mbps mobile connection, am I still a target customer?

If they have spare bandwidth they could actually undercut existing providers. And why not? The marginal cost of adding an additional subscriber is effectively zero. So the cost could change by location depending on how much competition for bandwidth there is. All they need to do is make enough money to increase the number of satellites and maintain the network. And as the number increases the footprint can be reduced. The satellite targets a smaller area on the ground. And you add more subscribers using the same logic.

It's an amazing business model really. You have a different price and a direct kind of customer in different places. A satellite starts an orbit over Antarctica and uploads scientific data. Then heads out over the sea and connects a few ships. Over Africa it connects rural 4g masts, schools and thousands of small businesses. Then over the Mediterranean it serves Netflix to cruiseline passengers. Over Europe it connects data for a NATO exercise and rural communities. Then later over London it sends data for high frequency traders. Then heads out over the North Sea to send YouTube to oil rig workers and telemetry for wind turbines. Heading up into the Arctic circle it connects a stream of airline flights to the web. And all you have to do is be slightly cheaper than the competition to win customers.

Thanks for sharing this, it's an amazingly unfortunate consequence of this technology which I never thought about.

I wonder how long people are going to be ok with these projects when one of them comes from China or Russia. Or even Germany...

Too bad the FCC does not think about space pollution. The astronomers were already very fed up with the Starlink constellation interfering with their observations, now they will have to deal with twice the number of satellites. Not to mention the risk of collision in space which keep going up.

The FCC and other regulatory agencies have carefully considered the impact of these constellations on astronomy and the risk of collisions, and they have (correctly imo) decided that the benefits strongly outweigh the costs, especially with the significant work these companies have done to minimize the adverse effects. If you have a particular issue with the principles the FCC used, or an objection to their computations, then please share that.

Quite wild, an American only agency/comission can say - "yeah, put 3,236 more sattelites in low-Earth orbit" and ruin the sky for the rest of the world. Do we not have an international body that would decide upon such rulings? With how fast this will escalate, we'll need to have one soon.

I’m pretty sure we do. The issue will boil down to who should control (or be controlled) by it. I can imagine resistance from say NASA to have it’s policies determined by nations physically incapable of reaching space.

And the outcome of the debate will be a function of your values versus someone else’s, which could then spiral into whether or not one persons desire for a nice view of the night time sky is more important than someone else’s access to affordable high quality internet (which is increasingly being viewed as a “right” these days).

Who knows how this all plays out but my gut says that we are observing oligopolies in the making. The first 2-5 will get licenses and then no one else, ever again.

If you go down the route of something being a right, but having oligopolies is _really_ bad, then it would seem that we should globalize space based internet. And why not? Civilization could transform in amazing ways if you can form a digital stream between any two locations on earth.

We already globalized highly accurate time and location, why not data streams?

Location is dominated by the US via the GPS system which our government has unilateral control over (hence competing constellations).

Part of me also sees a potential National Security security angle to an oligopoly of US-based companies providing satellite internet everywhere. Countries are increasingly putting kill switches on their internet. In a period of conflict, where information flow is likely to be highly managed, its convenient to bypass all that infrastructure with satellite internet. I’m sure these satellites can pick and choose their frequency. No doubt we could point these things at North Korea.

> Do we not have an international body that would decide upon such rulings?

The International Telecommunications Union comes closest.

>FCC does not think about space pollution

Orbital debris mitigation is an essential part of an application sent to the FCC

Things they care about:

* Will your satellites survive an uncontrolled re-entry? What parts of the satellite won't burn up in the atmosphere? Will any surviving debris have more than 15 joules of kinetic energy when it reaches the ground? If so, what are the odds of injuring a human?

* How will your satellites deorbit? If the satellite fails before you can deorbit it, how long will it take to deorbit naturally?

* How will you avoid collisions with other satellites?

etc. It's true that the process doesn't take into account light pollution, but that's just one of the many forms of pollution to worry about

I’m supposed to “think of the astronomers” when billions of people will get cheaper internet and advance humanity? Astronomers can find ways around this, we can’t stop progress for 0.0000000001% of the population.

The night sky isn't only available to some elite group of people. Although it's certainly difficult to see in cities due to the extremely high levels of light pollution. I recall an anecdote about power going out in a major city and residents calling the police up to report bright lights in the sky they had never seen before.

We should be trying to control our light pollution and space litter instead of increasing it, however. Seeing the galaxy with naked eyes shouldn't be alien to everyday people.

The reality is even quite poor countries are able to secure high quality internet for reasonable cost with appropriate policy. This isn't an otherwise unsolvable problem.

Thinking about space litter is a long term thing for other satellites as well. There comes a point where it starts making harder to guarantee that other satellites won't be affected.

there are already many satellite internet companies. Where are the billions you are talking about?

Are you entitled to clean air when 0.000000000001% of the population can make billions from selling fossil fuels ?

Air pollution results in a higher death rate and is a public health issue. Light pollution is certainly a problem, but probably not on the same scale.

This point I'm making here is that it's really just a matter of perspectives.

Even if one might quibble with the premise, parent's point is that the fortunes of the many outweigh those of the few. Your point is what... the converse? I don't think I can agree.

That is only because it’s the view of some peope this will be of net benefit to the world, flossing over some of the more obvious problems we are seeing the Internet cause today. Surveillance, disinformation, interference in democratic process to name a few.

On the other hand, soon space will be so accessible that any serious astronomer will be able to deploy a cubesat to assist in imaging and observation.

But I agree that this is an interesting dilemma. Perhaps the low earth orbit satellites should be limited to certain bands around the globe (certain coverage regions in the case of Starlink), so that there remain pristine skies in more remote areas: internationally agreed upon "Natural Sky Preserves" or something of that nature.

> so that there remain pristine skies in more remote areas

These pristine skies do not exist. There are already many satellites in polar orbits that intentionally pass over all points on Earth (e.g. for observation, or for the Iridium satellite phones). If you spend an hour stargazing in a recliner during reasonable viewing conditions, you will likely see a few passing overhead in the north-south direction.

A 15 megapixel smarphone camera on a cubesat is not going to be comparable with ground based imaging

A cubesat to replace an enormous ground based telescope?

This photo was doing the rounds a few days back showing starlink interfering with ground-based space photography: https://twitter.com/djulik/status/1286053695956881409

Also: https://en.wikipedia.org/wiki/Space_debris

The photo was disingenuous. Astronomers stack photos to remove interference. The photographer did the exact opposite.

I don't know about how the photo was originally presented, but the tweet that I linked explicitly says the frame is "17 30-second images of the comet added up", so it's hardly disingenuous.

The impact of Starlink on astronomical observation is in dispute and by no means is a settled matter: https://en.wikipedia.org/wiki/Starlink#Criticism

He's saying the photographer specifically stacked the frame to increase the interference of the satellites. The same basic technique can be used in either direction - astronomers already use it to remove the thousands of satellites and other causes of interference in the night sky.

https://www.youtube.com/watch?v=mgMZa6G-q34 - a basic example using it to remove ISO noise

Edit: Actually, you can even see some satellites in the shot when he sets the layers to 'lighten' that are totally gone when he finishes the stacking.

Isnt this just using long exposure?

No. A single long exposure would only show maybe one or two of those satellites. That's numerous exposures stacked on top of each other, such that every satellite visible in any of them is in the final result. This is completely backwards from how photo stacking is meant to be done (it's meant to remove satellites not visible in all of them.)

Very interesting, thanks for sharing this. This indeed should be pointed out clearky by whomever creates such images - seems quite manipulative using this to make a point.

These space internet monopolies look increasingly like an international security and privacy nightmare in the making. Space wars indeed and i wonder how non US/Korea/china people will be able to protect themselves.

Applications are open for YC Winter 2021

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact