Starlink is definitely filling the niche market of providing internet connectivity in rural and sparsely populated areas where firms with traditional business models with fiber-optics based network are not incentivized to provide. The real question is: will Starlink become profitable by in turn finding a big enough market itself to be able to cover the Billions of dollars of cost of manufacturing, launching and maintaining these satellites? Elon Musk seems to be open to separate Starlink business from SpaceX and list it publicly in future (he tweeted something in that regard previously [1]) so it seems like he definitely thinks Starlink can become a profitable business in near future.
On the "big enough market" question, passenger aircraft, and commercial shipping, and US military make for a compelling business case in of themselves. Each of these use cases have customers that will pay over the odds for a reliable connection to remote areas.
This guys thread will almost certainly be wrong and look stupid in the long term. Pessimistic predictions sound smart in the present because with anything new it's hard to know how it will play out. Long term though any early issues with new technology tend to be optimized or solved as things evolve.
>>Long term though any early issues with new technology tend to be optimized or solved as things evolve.
I'm absolutely, completely, utterly sure someone said this about the first fossil fuel technologies. "Oh, I'm sure this 'pollution' thing will be solved."
I think it's too early to tell, and it depends (a lot) on how the next ~25 years go. Probably not though, but that's mostly because both a) the world is so much better than it was pre-industrial revolution and b) I don't think that leap was possible without fossil fuels. It's not because pollution/CO2 turned out to not be a problem, quite the opposite.
With Starlink - offering cheap and high-speed internet access to currently underserved places is absolutely good and important, maybe even critical for long-term success and stability of some regions. It's not clear that this is the best way to do that. Kessler syndrome, light pollution, potential upper-atmosphere effects (ozone depletion), should not just be waved away. It's just more externalities piling up. Rich countries should be subsidizing these underserved areas around the world so they can put in infrastructure - cell stations, fiber lines, etc. - rather than filling up the sky.
Take a quick look at the economic situations of the world pre-Industrialization.
Most riches we experience in this modern life... depended utterly on the energy credits we spent by burning the eons-old reservoirs of solar energy, of dead plants and animals.
And the biggest catastrophes we will experience in this modern life... are almost exclusively due to the energy credits we spent by burning the eons-old reservoirs of solar energy, of dead plants and animals.
Honestly not sure it's a win. Maybe they had less, but they had not destroyed most trees, mammals and insects (which we have already), and all that for what? For a future where we will most likely have much, much less energy, and therefore where most our modern system... will collapse.
Others have pointed out things wrong with that doomsday predicting Twitter thread. I wanted to add a couple of things
* Rocket launches are a rounding error on co2 emissions when compared to passenger aircraft
* the worries about Kessler syndrome whilst simultaneously worrying about tonnes of aluminum and rare earth minerals (gasp) due to the short orbital lifetime of the satellites are contradictory.
* Presumably astronomers are more worried about light pollution rather than ecological concerns
On a side note, what is it about people hyping up rare earth materials? The twitter thread tries to make them sound scary. Others try to make them sound harder to find than gold.
On the light pollution point, twilight is the worst time for satellite brightness, but as the satellites enter the Earth's shadow their brightness rapidly diminishes. Some estimates are 10%-30% data loss during twilight for Vera Rubin observatory and 2% for vlt. A problem for Vera Rubin, but not so much for other observatories.
Naive eh? What "nasty things" do you think are released into "every layer of atmosphere"?
Since we're specifically talking about SpaceX and Starlink, let's talk about Falcon and StarShip rockets. Let's look at the "long list of very nasty things".
Falcon 9 is expecting to launch a record number of flights this year, recent reports suggesting 52. Falcon 9 burns RP1 (a refined kerosene, very similar to jet fuel), a hydrocarbon that mostly produces CO2 when burnt, but also Carbon monoxide, and plain old carbon. The Falcon 9 has about 200k litres of fuel when fully loaded. It sounds like a lot, but for perspective a fully loaded Airbus A380 has a fuel capacity of 380k litres. So, they output of a record breaking year will be the equivalent of around 25 long distance flights of a large passenger aircraft. I'm not going to find out how many long distance large plan flights are happening, but many estimates are over 100k a day. So a record breaking year outputs a fraction of a percentage that airline flights do each hour. And the "nasty" stuff is mostly CO2, CO, and C.
Starship will burn Methane, and will burn even cleaner than Falcon 9. It will also be fully reusable, resulting in less waste again. On the downside, methane is a potent greenhouse gas, so any unburt fuel would be not great. That said, rockets are designed to make the most efficient of fuel, and any unburt fuel is not going to be released in normal operation.
> I'm no rocket engineer
Colour me shocked!
As for going through "every layer" of the atmosphere, I wouldn't get too worried about that. A lot of a rocket's fuel is burnt at lower altitudes (the same level at which passenger and military Jets operate) and the various zone mix readily, and are really so massive that you would have trouble measuring the impact of hundreds to thousands of flights of Starship per year.
> When it comes to greenhouse gasses alone, the contribution of rocket launches is rather negligible, accounting for only 1% of aviation's carbon footprint — which in itself only amounts to 2.4% of annual global carbon emissions, according to the Environmental and Energy Study Institute
Apart from the fact they overinflated the impact to 1% (there is no way it is close to 1%) there is a reason they chose the adjective "negligible" to describe the impact.
> Eventually, the circulation of the atmosphere dissipates those greenhouse gases and brings the concentrations back to "standard" levels. It is, however, unknown how long the concentrations remain elevated and what their impact on the temperature of the mesosphere could be.
Reality is that these scientists did a simulation, got some results that can be made to sound alarming, but just don't know if these is any inpact. I'm all for studying this more to see if there is a real issue. My gut says it's not going to be a problem until we have at least a couplr of orders of magnitude increase in launches, but it's worth looking at it further.
And I'm not sure why you included the graph of objects in orbit? This looks exciting, but it really is just that companies are now launching clusters of smaller satellites. The good news is that these tend to deorbit in under ten years. Unlike practices of old, this new style of satellite will not result in long lasting space junk.
Finally, what is the point of including a graph of airline flights over time? What bearing does it have on rocket launches?
I read an opinion piece somewhere that Starlink is just busy work to keep SpaceX launching rockets, but because gestures at Google I can’t find it. This Twitter thread nakes this theory plausible.
Starlink is absolutely a way to keep SpaceX launching rockets, but that doesn't make it busywork. It's revolutionary and immensely valuable.
And, any theorized environmental impact has to be compared to the existing natural process of space dust falling into the atmosphere, and measured against the saved costs worldwide of wires and cell towers and travel not needed, especially as capacity increases in the future. Which is practically impossible to measure, but failing to account for difficult-to-measure costs is a classic error people make when evaluating things like this. And of course there's the extra benefits of coverage over oceans and poles and to ships and planes that you can't get any other way.
Starlink directly competes with wired internet. It also competes directly with cell towers in the fixed wireless business and in mobile service for vehicles. Next year Starlink will be competing for mobile service direct to handsets too. Sure, people will use more bandwidth when they get Starlink, because Starlink provides more bandwidth. But that won't increase demand for Starlink's competitors, all it means is people get better service. I find the idea that Starlink doesn't reduce demand for its direct competitors extremely implausible.
It doesn't make sense to lose money on every rocket launch but make it up in volume. It's more accurate to say that reusable rockets and a LEO satellite constellation can form a mutually-reinforcing flywheel that will lower costs for the eventual Martian fleet.
Why can’t it be both? It absorbs excess capacity, allowing spacex to expand their capacity and achieve economies of scale that are unmatched in the modern world of rocketry while also setting up an exciting new revenue stream that will move them away from relying on government contracts. It also helps them win support from republicans as it serves rural communities and is an example of business innovation that the GOP can support which should help them win more govt contracts. Lots of win-win-win opportunities here
This thread is mostly making large conclusions from limited study at google. Lots of it isn't really that true.
And he makes it quite clear with 'Not interest in discussion'. His mind is made up from when he was at google and that's it.
The industry isn't the same as it was then and it was mostly economical that google didn't do the project.
The ecological isn't really bad, its mostly that rocket launches produce CO2, but for a global infrastructure project it isn't actually that much. Everything else isn't really that relevant.
Fully and rapidly reusable rockets, like Starship, if achievable, will change everything. I don’t think peopel comprehend how the economics of launch will change.
Satellite breaks down or runs out of propellant? Send up a repair/refuel/reentry vehicle.
Astronomers upset about light pollution? Launch a space telescope for them for free, and eliminate that pesky atmosphere for them too.
The usual form of the metaphor is "a crab bucket".
The mental image this is meant to evoke is that if one crab attempts to escape the bucket by climbing the side, then others will grab it, and their weight will pull it back down. So no crab escapes the bucket.
A metaphor for unconstructive criticism or obstruction that squashes attempts at betterment.
If it needs explaining I meant this guy is worried about a few 1000 tons of space debris versus the certain eventual end of the species if we don’t become spacefaring.
What about we start by not destructing our planet? "No no but look: we destroyed most life on a whole planet, but it will be different in a spaceship: there we will live happily in peace forever, without any problem we can't handle (because we destroyed them with our habitat)"
This is not what the grandparent meant, and also not what anyone sane means when discussing establishing off world colonies. They are talking about an insurance policy if we happen to get hit by an asteroid, or world war III breaks out, or a devastating pandemic, or any number of other civilization (or life) ending events.
No one wants the bad thing to happen, but instead they are saying "wouldn't it be smart to have redundancy". They're also not saying "ignore trying to save the environment because we have mars". It is very likely any effort to make mars into a habitable planet would dwarf fixing problems here on earth.
They're also not suggesting spending a significant amount of our resources on Mars, to the detriment of life here on earth.
That's the thing: not everyone agrees on all that. I personally don't give a damn about saving a subset of humanity by sending them on Mars. But by doing that, we pollute, and that's part of the problem here on Earth (which I care about). Worse: they're not just sending a couple rockets, they send as many as they can, and build services like Starlink, which are about producing and consuming more, which again is part of the problem.
I am of those who believe that if technology is a big part of the problem, then more of it won't make it better. But it's not like I have a choice, so instead of waiting for a potential asteroid that will kill me while some (rich US citizen, maybe?) survive on Mars, I am here waiting for the catastrophes we are creating.
The bigger question is will Starlink be able to retain the customers that they have now as more FTTH is deployed deeper into rural communities. My own company's FTTH service is: faster (1Gbps), better (lower latency, no congestion issues) and cheaper (costs from CAD$30-85 per month less than Starlink). Meanwhile, the incumbent here has just launched 8Gbps service. Starlink is better than what came before, but it's not better than fibre at present.
I thought the move to do midband mobile spectrum with t-mobile and offer line of site coverage across 100% of the country was a pretty big move. Existing phones can be enabled on this spectrum. I also see it being leveraged well for edge services where classical 5g towers are just too damn expensive to pout out that far
You do realize that in terms of speed it makes 2G look like a dream?
It’s only good for emergencies.
And anyway it was mostly just a publicity stunt, announcing non existing feature (no satellites in orbit are able to do it) to preempt Apple by few days, and them rolling out working product this year.
"The public tends to respond to precedents and superlatives," is one of his main quotes. At some point I think he figured out it doesn't even need to be released first to get the precedent credit.
I'm a Starlink customer. I regularly get 100+ mbps in Austin, TX and have not noticed any speed drops. If anything, it's gotten more consistently fast.
I just setup a wireless bridge this weekend for a farm that needed to stretch their starlink.
I found some interesting behaviour on start link..
Using the starlink in built speed test app I would easily get 150mbps
using the ookla [1] speed test on a desktop machine I'd get no more than 55mbps and on a mobile phone I'd consistently get 20mbps slower than the desktop.
ookla tells you where you're getting your speed test sent to but the starlink one doesn't... is it a speed test just to the satellite in LEO?
I always use https://fast.com/ for speedtests, it's hosted by Netflix (I rarely use the Starlink app after initial setup). I checked just now and clocked 71 Mbps, 160 Mbps, and 120 Mbps on 3 different runs.
Running the Starlink app, it's clocking 108 Mbps (within same range reported by Netflix).
Just tried speedtest.net and clocked 152 Mbps.
Sometimes speed can drop temporarily when satellites fall out of range, so I would try multiple times over a 2-3 minute window to get a more accurate average.
Underserved areas are like 98% of the earth’s surface and include some rather wealthy entities like airlines, shipping/logistics, mining and resources, cruise liners, billionaire’s yachts, and of course governments of small/island countries.
And all of that adds to maybe 100k customers. What next? Especially for a service that has scale bandwidth everywhere across the world, it increase it in one spot (or add 5G tower there, which you can just do, without this whole “10k satellites that have to be replaced every few years” business).
I’m not saying there’s no market. But it’s very limited. Much more limited compared to all the hype it’s getting. I had literally people living in SF telling me how they’ll use it, to stop dealing with Comcast.
They literally already have far more then that and they have only just started. They don't even have full coverage yet and have only just launched their mobile products and only just started their shipping products.
Airplane and shipping industry isn't even really onboard yet, there is a huge costumer base there.
The military, NASA and many other government organizations, has lots of uses for it as well. And that goes for other nations equivalence of those organizations as well (except maybe military depending on US policy on that).
> “10k satellites that have to be replaced every few years” business
Every generation the sats will get vastly better thus increasing the capacity and/or decreasing the price. At the same time the rocket technology gets vastly better decreasing the price considerably. This first iteration is by far the most expensive one.
Those 100k customers are not like the others. Commercial, government, and military customers will pay FAR more per connection than an everyday customer.
This isn’t even getting into backhaul, assuming the laser links work out. There are financial firms that would hand over untold stacks of cash for even a few ms of performance improvement between say, London and NYC. And Starlink can do that between all sorts of distant city pairs.
Surprising absolutely no one who understands networking.
The ludicrously high speeds that were praised during the limited beta really seem like they were designed to be a bait and switch to get people on board with completely unrealistic expectations for a full fledged satellite internet service.
I have Starlink and have been using it since last year. It’s true, the speeds have gone down during peak hours. I get maybe 100Mbps instead of 250-300. And I see speeds back in the 200s off-peak.
But damn, Bait and Switch? Harsh words for extremely usable high speed Internet from Space that’s fast enough for remote working, online games, zoom, and 4K streaming that works extremely well from a good portion of the Earth. Compared to “satellite internet” of the past it’s a similar technological leap as dial up was to cable internet. I don’t know how people can complain.
It seems to be surprising Starlink; they're still telling customers they should expect speeds far above what we're getting in practice. Their service description is a fiction.
You make it sound like they are being extremely misleading, and yet the page you linked to has this text in bold:
> Stated speeds and uninterrupted use of Services are not guaranteed. Actual speeds will likely be lower than the maximum speeds during times of high usage. Starlink may temporarily reduce speeds if our network is congested.
They've adjusted that language (and target speeds) several times without altering the price in the US. Also it's deceptive to publish a bunch of speeds and then say "lol just kidding you won't get that".
Of all the numbers the most fictitious is the 20-40ms latency. I've averaged 55ms for the last month and it's basically been that way since February 2022. The variance is wild too, 25-150ms is common.
At least starlink publishes expected service speeds somewhere [1]
Xfinity provides "800-1200 Mbps service" in many areas and they don't even mention upload speeds. You can expect less than 50 Mbps for areas, though their official press will advertise "eligible xFi Complete customers can receive up to 5–10x faster upload speeds" [2].
Yes its amazing that a company that is deploying more infrastructure every week thinks that the capacity of their infrastructure will go up. Crazy that they think that.
I wonder how expensive they are to produce now and have they managed to break even. Gwen Shotwell told the press that their costs were around 1300 with some updated model that was released in early 2021, she said that in April. Since then there was second gen antenna made, that is, I presume, cheaper to produce. And there was leaked email of Musk in late 2021 that was talking about need to get v2 sats working so amount of terminals they are ordering could make sense(email was about their need to fix Raptor 2 production in order to do this). Then they upped their prices in March of this year and IMO this could be a sign of break even point(not sure though).
I’d guess that the fact that they are allowing service you can pause now means that they are probably at break even on the user terminal hardware cost.
1 million terminals shipped means roughly $1 billion a year in revenue. Some customers are paying substantially more than $100 a month depending on the tier of service.
If one assumes that all terminals are in use by paying customers @ 110[1] per month that gives 1.3 billion a year in income. Plus 0,5 billion usd for the hardware.
So far there have been 62 Starlink launches - assuming an average price of 67m usd[2] that gives a total launch cost of 4.15 billion(real launch cost is lower as SpaceX does it at cost.
Life expectancy of the satellites are between 5 - 7 years so excluding R&D cost I'd say Starlink is already cashflow positive and will generate huge returns in the future. I'm sure there are some commercial / military contracts in the mix which alone will add a couple of B's to the income number.
Your numbers are a bit off, it's sort of safe to assume that maybe half of terminals are operational - not 100%.
> some commercial / military contracts in the mix
T-Mobile is a massive $$$ for them.
I would assume the availability of worldwide, omnipresent, inexpensive low speed broadband is going to open new use-cases (and thus revenue streams), that we're not even aware of. It's going to be exciting.
> Your numbers are a bit off, it's sort of safe to assume that maybe half of terminals are operational - not 100%.
Wait, why is this? Do you mean that 50% have been produced and not yet sold, or 50% have been sold to people who aren't paying for service? I'd assume a much smaller percentage in either case.
I think the current estimate for the direct cost of reusing a Falcon 9 is at most $30m and potentially as low as $15m (there's a lot of fixed costs so getting a "true" price for this would be hard even if you had access to the same data SpaceX does).
Once the satellites are up there orbiting, they might as well be earning income for as much of each orbit as they can. Zero marginal cost to do so. So SpaceX's prices in each market will be low enough to be competitive in each market.
The caveat to that is that they're still taking a large loss on each user terminal, so they have to make that back over the expected life of a contract.
1m user terminals manufactured and yet we still don't have one 18 months after paying for it. (Actually, I canceled my order a year ago. But my brother, who also paid the full $600 still doesn't have his. And they keep delaying it.)
Three years ago we both signed up for a terminal, paid the $100 deposit and got on the list. Eighteen months ago I got an email saying I should log in, pay the additional $500 to have the terminal shipped to me. After three months of waiting nothing had shown up. There's no way to contact customer support so I had to ask a friend who works at SpaceX who I should talk to. Two weeks later I got an email from someone I'm guessing is part of their support group saying i was in the queue for delivery in late 2021. Late 2021 comes and goes and no terminal. I ping that same person. Now I'm in the queue for a late 2023 delivery. I ask if I can get a refund, they say yes. They credit my card for the second $500 charge but not the original $100 which they claim is unrefundable.
My brother, the true believer, is still waiting for his terminal and his cybertruck.
I know that Elon is a minor deity on hacker news and peeps are going to downvote the heck out of this comment, but from my perspective, dude's a ripoff artist who survives only because the government is paying for his rockets and ev subsidies.
You’re in an oversubscribed area, it’s that simple. That’s not going to change until they either get more sats up there or people in your area cancel service.
If you email me to tell me "pay for your terminal and we'll ship it to you within 2 weeks" and instead of shipping it to me in two weeks you tell me, "thanks for the cash, we decided to send it to you in 18 months," then the problem is not that I'm in an oversubscribed area. The problem is SpaceX is incompetent when it comes to basic logistics or is less than truthful regarding its service.
It’s been my thinking for a while that the majority of people who interface with Musk ventures walk away happy. They must, right? Customers are the most important fans.
But if for some reason one of the cogs in the big machine fails you - you are screwed. There is no real recovery beyond tweeting @ElonMusk on the day once a year has his heart reinserted to legally be considered alive.
I’m sorry to hear your situation - mine was different, more positive.
The line that separates us is one broken cog. Musk customer support is the worst and I’m sorry you can’t enjoy what has been a life changing product for me.
Hang in there, it (Musk venture customer service) sometimes gets better.
Musk quote: "We specialize at turning the impossible into late."
Personally, I much prefer the latter. I also quite like the transparency that lets us all see into the disorganized mess of it all long before it's ready (as opposed to the Apple strategy of complete silence until announcement day).
I think it's more that your comment is needlessly abrasive rather than anything else. And I don't think it's a healthy mindset to say you thrive on toxicity.
Would love to see websites start optimizing for low-bandwidth high-latency connections. Page sizes have continue to increase even with the increase in usage of [typically] slower mobile connections, so I'm not holding my breath.
One thing that has always been true is that we always put as much crap as possible while not getting unusable.
Computers got infinitely more powerful, but websites are not faster to load because they are much bigger. QUIC won't change that, it would just allow for more crap.
You don’t need to go rural or sparsely populated for low cost living… places like india have low cost housing, food, internet, and labor. I would think the market for internet in sparsely populated areas is by definition negligible.
But, if all cell phones can access starlink in cell phone dead zones, that might make it economically viable, as the market expands to 5 billion plus users.
The profitability of the placing and maintaining physical network infrastructure (cell towers etc) in ”dead zones” is low. The profitability of covering many “dead zones” with a single orbit may be different.
Perhaps. But another way to look at it is, now you could live anywhere. Pardon the cliche but Starlink might a game changer, a paradigm shift. I don't disagree with you per se. But in this case the future might be different than the past. We'll have to wait and see.
That’s some funky reasoning. Most Americans will not move overseas to get lower cost of living, so rural areas are a good substitute. Likewise, cost is not the only consideration, also quiet and privacy.
Not something to be proud of when you have less half a million customers. Literally half the terminals are e-waste because you can't pause service or pass the terminal on to someone else to use.
.png){kind=link}
[1] https://www.cnbc.com/2022/06/07/spacex-starlink-ipo-elon-mus...