DEBUNKING STARLINK

Yet another Elon Musk project that doesn't live up to the promises - and this one might be the worst one yet!

8/10/22 UPDATE:
StarLink failed to meet the basic requirements for the $855million FCC grant they were counting on.

[ian_b]

I think you're very unfair to the visionary genius who invented slow underground taxis.

[AndreiBarsan]

For the record, ping is not exclusively a gamer thing. Most aspects of remote work suffer under high ping---zoom calls become difficult with people talking over each other, remote code editing is difficult, remotely interacting with servers is annoying (SSH with 1s ping, anyone?), collaborative document editing gets difficult, etc.

[tjakal]

Whenever I catch myself questioning if leadership could really be this deluded I'm reminded of that congressman who wanted to task the Forest Service with altering the orbit of the moon.

[LifesLaboratory]

Starlink was never meant for customers that have a fixed line available. I would gladly have fiber or cable if I could, but I'm in a rural area of central BC. I consistently get 90mb down and 8mb up on Starlink. It has never gone down for more than a minute in the nearly three years I've had it, and this is a very rare event. Other than Starlink, the only option for me was cellular, which I had for eight months. With cellular I got less than 5mb down and 0.5mb up. The cellular would often go out for an hour or more... sometimes a full day! Starlink has far surpassed expectations for me. For many people (like myself), Starlink is a necessity for working from home. There are enough people on the planet that will gladly pay for this service to make it profitable.

[michaelfuchs]

"It's not that difficult. Think of it as an thermosphere - with an airhockey table!"

[derkach]

I found a legitimate business idea:
Buy the dishes for 500$, sell them to spacex for a 1000$. Repeat.

[jslabotsky]

Some good math here, but I want to share some thoughts as someone who operates a fixed-wireless ISP serving rural customers whose only other option is satellite. In the section about data rates, I somewhat take exception to the way you characterize residential users' bandwidth requirements. Almost nobody needs as much bandwidth as they think they do.

During our prime usage period in the evenings, when everyone is streaming, average bandwidth consumption is 4.5 Mbps per user. I was surprised it is that low, but I have the receipts. Every WISP I've spoken with has reported their usage is right around the same number. We tend to assume more bandwidth is always better, but the reality is the vast majority of users reach a point of diminishing returns somewhere around 20 Mbps. Above 30 Mbps, an increase rarely has any noticeable impact on user experience. Latency, packet loss and uptime are far, far more important to user experience than overall capacity. The point being, the price vs. bandwidth comparison does not tell you anything about the user experience.

Here's a real world example: My business partner has both our service and Cox cable at his home. Cox can, theoretically, provide him 960 Mbps of bandwidth. Our WISP can run about 400 Mbps. But Cox's packet loss is so outrageous, and brief outages are so frequent, that it is only useful as a backup. He gets way better metrics on our network, so our network actually provides the better experience despite providing half the capacity.

With that said, I know it is pointless to try and convince anyone they might not actually *need* gigabit. So I accept that it's a losing argument, if not a particularly logical one. Moving on...

Another thing you never touched on with regard to bandwidth is frequency spectrum. SpaceX has licensed quite a bit of spectrum, but it still is a finite resource. This is what will limit aggregate capacity in heavily populated areas. Radios have to take turns talking to the access point. You can only transmit so much during each turn. Retries eat up more turns. The larger the number of clients, the longer you have to wait before it's your turn again. As you mentioned, the bigger you scale, the more individual performance suffers Scale in wireless of ANY kind will always be the biggest challenge. It will be no different for Starlink than it has been for WISPs for a couple of decades.

THEN there is the fact that most of the spectrum Starlink has acquired is very susceptible to atmospheric attenuation. Having looked at link budgets for 11 GHz and millimeter wave radios myself, I'm having a really hard time seeing how they're not going to be impacted by weather considering the size and power constraints of the subscriber equipment.

Overall, functionally I think Starlink could be a great thing for their original stated purpose of reaching the most remote parts of the world AND being able to support applications that need low latency. It could open the door to new applications in medicine and science that improve people's lives. Fiscally, however, I don't see how it will ever be profitable as a residential service. Military and commercial applications, on the other hand, could easily pay the bills. So I wouldn't write them off just yet. But I do need to see more before I would invest. For consumers, there are better ways to improve access to broadband.

One last thing, if anyone has made it this far - as a business owner, there is NO WAY I would use subscriber equipment that has a 15 MONTH ROI before that customer's revenue contributes one penny to paying operating expenses. If I can't make that money back in the first 2-3 months, I'd be out of business. I know Musk says they're working on reducing cost of the subscriber unit, but it's going to require some kind of quantum leap of engineering to pull that off. Basically, the business model doesn't work without government subsidy. Instead of giving $900 million or so to SpaceX, it would have been far better to award it to small, regional WISPs who want to get into fiber or to existing small FTTH providers. These small companies with a vested interest in serving their communities could get fiber to millions of customers and leave behind something with real longevity, less ongoing cost and maintenance, and far less negative environmental impact. Unfortunately, the way RDOF and CAF funding works, most of the smaller companies can't even afford to file the paperwork. I'm telling you, there is an army of small, locally owned providers who could do a kick ass job bringing fiber to under-served areas if they received the same kind of funding, proportionally, that SpaceX and big telecom do.

[LouigiVerona]

Should there be an update on this? How is Starlink doing?

[FlintSparkedStudios]

I’ve gone from “I don’t like Elon Musk but I like the projects he’s doing” to “Wow the richest man in the world is a con artist.”

[someguy]

If people in third world could afford $100/month on internet, telecom companies would be rushing in to provide internet to the remotest corners of the country.

[jerbear]

If I was taking a loss on units I would imagine I would want them to be pretty easy to fix.

[aether]

I would love to see this revisited. I live in one of many rural fishing villages in Alaska where the environmentalists will never allow lines to be run to us. Because there was zero competition, the local Internet company is a crime in progress. $350/mo for laggy YouTube surfing. Starlink was $500 down and $90/mo for unlimited data and 99% less lag. I've gotten my money out of it already, but I really hope it doesn't go away.

[Jf]

Starlink has been a game changer for me. I live in the middle of no where. Every previous internet provider was under 10mbps download and I was paying over $100 per month for a data cap of 500gb per month, it was so intermittent, sometimes I would have no internet service at all. Now I have over 100mbps and pay roughly the same. Zero outages to date. Pretty happy customer.

For those who already live in city’s with fast & cheaper internet services. It would be a poor choice to switch to starlink in my opinion.

[nathanwhitman]

As someone living in a somewhat remote area and who's been trying for years to get decent interne, i have to say that Starlink is an order of magnatude better than any other service i've tried so far. I have tried Hugh's Net which on a good day will get maybe 15-30 mb/s download when not being throttled and less than 1 mb/s after being throttled . Some months I used my allotted data within a few days. LTE hotspot was around the same data limit with much better latency ... though peak download speeds were 5-10mb/s . Now on starlkink I have 50-200mb/s speeds, low latency, with NO DATA LIMIT. The connection is a bit unstable, but no less unstable than Hugh's Net or an LTE hotspot. Overall this service is a game changer for me!

[MehediHasan-dlle]

The soon to be "Founder of Twitter"

[bartzrt]

The criticism of the business case is really solid I think, but not the criticism of the service. If you live in a remote area, Starlink is way better than other satellite providers. You try browsing the internet or watching YouTube with a 500 ms ping, it's not a great experience. Having said that, your argument that there just aren't enough people who have both a need for this service and the money to pay for it is a good one. It's probably why SpaceX is kinda pivoting to more luxury applications (RVs, yachts, offshore...).

You need to stop with the "people in Africa are too poor to need internet" stuff though, it's incredibly condescending and doesn't reflect reality. Access to internet is often a massive boon to rural communities. Now, I don't think Starlink is all in all a good solution for such communities. Ground-based directional radio relay stations are much cheaper, easier to maintain and more reliable, plus they offer a clear upgrade path to terrestrial cable.

[Ironbattlemace]

You are pretty unfair against a man who stole a broken idea for vacuum trains and has working underground taxis with RGB lights.

[bccbuue]

As a former networking professional (I worked primarily on fiber optic transponder firmware) and current Amazon emplyee, I agree with almost everything said here. The only thing I disagree with is that only gamers care about ping. A high ping will definitely be noticeable in general web browsing, as modern web pages often make many requests to back end services, some of which rely on one another. This would mean that only many pages, the absolute minimum loading time is some multiple of your ping. At geostationary orbit, iirc, the minimum possible ping is over 200 ms (with many providers being several times this), which means that that is also the absolute minimum time for any web page to load (assuming infinite bandwidth). With a more modern webpage that is making many request to back end services, this time could easily double (with just one call depending on data from another) and would likely be much more. I remember when we pages took more than a second to load, and I really don't want to go back to it.

[MidNiteR]

Starlink, Solar City and the Hyperloop will make great episodes on Engineering Disasters on the history channel.

[brianbutton]

In October 2023, Wall Street Journal reports Starlink revenue up to $1.4B but no word on profitability. This is a slim 20% of revenue targets set several years back. The *'subscribers-per-satellite'* count is up to about 250, from 150 they recorded 18 months ago. And, *'revenue-per-satellite'* is $250K for 2022.

Those stats show steady improvement and deserve some respect. But I see no sign they are about to climb out of the impossible economics box they are in. I would think that 2, 500 *'subscribers-per-satellite'*, 10x their current number, would make for a sustainable company.

The cost of a manufactured satellite, still in the factory, is $250K. Launched and managed, it is a couple million. So they are already in the hole at their current *'revenue-per-satellite'* . _(Lose money on each one but make it up in volume.)_ And they have a fair ton of other expenses, overhead and operations.