Hacker News Reader

Comment by chroma | Hacker News Reader

If you look at the authorized bids[1] versus the defaulted bids[2], you can see only three companies got any money from the Rural Digital Opportunity Fund broadband subsidy. The other five companies got nothing, including SpaceX and LTD.

E-Fiber San Juan LLC gets $7.5M to provide broadband services to 1,085 homes in rural Utah. Monster Broadband gets $5.8M for 11,286 homes in Texas and Tennessee. Northern Arapaho Tribal Industries gets $7.8M for 2,408 homes in Wyoming (probably on Indian reservations). So a total of around $21 million will be spent on 15,000ish homes.

The FCC's budget for the RDOF subsidy is around $20 billion for 4 million homes, so 99.99% of the money is still up for grabs, as are 99.96% of the rural households.

1. https://docs.fcc.gov/public/attachments/DA-22-848A2.pdf

2. https://docs.fcc.gov/public/attachments/DA-22-848A3.pdf

View on HN

To be clear, those links are only for the eleventh batch of "ready to authorize" bids that was released today. Other bids were authorized in different batches, viewable at https://www.fcc.gov/auction/904/releases

I have a feeling that Starlink will end up being an indirect recipient despite the rejection.

Starlink might just end up making deals with the companies that won their bids and sell their services/hardware as rebranded solutions. It's probably cheaper for these companies to do this instead of setting up their own infrastructure.

Of course the real question is probably whether not following through on their obligations and getting fined is less than the amount they were awarded.

Starlink won in the 100Mbps category. Almost everybody else won in the 1Gbps category, so can't sub out to Starlink.

The way the rules were set up, virtually any bid in the gigabit category beat the megabit category, so Starlink only won in the regions nobody bid in.

Does that reflect permanent limitations on Starlink speed? Somehow that's less impressive than it initially seemed...

Yes. You will likely see a maximum theoretical speed of about 600mbps (like WiFi). In practice it will be much slower, maybe half. It’s noise shooting through atmosphere, so there will be physical limitations.

Wireless will always be worse than proper fiber.

In a perfect world, we would all have active optical fiber from the backbone to our home routers. But 5G and satellite internet is cheaper...

Cheaper, in some situations. E.g. starlink would be horrendously expensive to provide service to everyone in downtown SF.

But yes. For rural less dense deployments. Long range signal through the air is much cheaper.

Agreed, problem is with everything SpaceX hard numbers are hard to come buy. So wireless will be cheaper, if satelites are cheaper than 5g and other groind based solutuons is hard to tell.

Cell towers already provide internet to everyone in downtown SF and have for 20 years. Just because you aren't using your PC with it normally doesn't mean it doesn't count.

Wireless is always cheaper than wired for an equivalent coverage area. It just requires less physical infrastructure. But in almost every other way it has inferior performance.

Could you explain your reasoning? My understanding is that:

fiber=c lasers=c fiber=nonlinear lasers=linear

So in physics terms, starlink will eventually surpass buried fiber in latency for all situations where fiber routing divergence from linear exceeds the ~1000km orbital round trip.

Now compare the noise level in a buried fiber, versus satellite radio transmission.

We've been getting a thunderstorm every evening here, and my Starlink slows down a lot -- not sure if it's ionization, the water content of clouds, or what, but it's definitely noticeable.

And my previous point-to-point wifi would also essentially stop transferring data when it rained, just in case some Starlink hater thinks it's somehow worse than alternatives.

Technically the speed of light in fiber is 2/3c because fiber has a higher index of refraction. Some HFT funds have actually moved away from fiber in some instances because they can pick up a few microseconds by going through the air instead.

Goldman Sachs actually moved to a neutrino beam between London and NYC.

there's nothing about this on Google, and neutrino detectors are definitely not advanced enough for this. was this a joke that went over my head?

Maybe it's a joke about the (anomalous) detection of faster-than-light neutrinos?

https://en.wikipedia.org/wiki/Faster-than-light_neutrino_ano...

I can't find anything about that online is this a joke or is this actually real

Starlink will get slower and slower as they get more subscribers.

This is already happening. In many places it is well below half as fast as before. More, and more powerful, birds will help some, but if they need to loft 4000 a year their unit cost will matter.

I expect they will cap subscriptions offered in some areas, particularly parts of Asia and Africa.

True, they aren't upgrading the satellites to the v2 ones that are 10x more powerful in total throughput and number of subscribers. They also aren't launching more satellites (and rockets) every year /s

Oh wait, I was wrong on all of those!

Don’t the next gen satellites require the new Starship for deployment, which is still nowhere near ready for launch (they aim for a first test launch by the end of this year, IIRC)?

They have launched exactly zero of those "more powerful" satellites. And they are not launching at near the replacement rate for the final constellation.

Higher-bandwidth birds just moves the bottleneck.

It appears you are doubting that they will eventually launch their 2.0 satellites with an order of magnitude more bandwidth. It's not like they haven't delivered in the past...

The constellation is still growing quickly, they just passed 3000 launched Starlink satellites. By the time they go past the 9.000 satellites or so they should have Starship ready and kick their launch volume into overdrive. Have the 30,000 extra satellites (on top of the 12,000 already approved ones) even been approved yet?

> Higher-bandwidth birds just moves the bottleneck.

What does that even mean? There is no reason to believe they can't launch enough powerful satellites to reach faster speeds in those rural areas.

Not getting it? Putting faster routers on the satellites has absolutely no effect on the radio bandwidth available between sat and terminal.

If each new satellite can put down more spot beams, you can cover more cells for longer. With the spectrum they have, they could put 8 beams on a cell with 100% duty cycle, which is probably around 3.2-4Gbps.

Use narrower beams and you increase the availble bandwidth. Same as using more wireless access points using lower transmission power each.

They’ve already applied to the FCC for the 2Ghz spectrum. They’re already launching the larger 1.5 satellites that include lasers to limit ground stations. They just had a successful test fire of a starship with the engines mounted this week.

What makes you doubt the company that made reusable rockets that land on boats and reflux 13+ times won’t do this? The ENTIRE industry said it was impossible but it was fully within the realm of physics. They pulled it off because it was possible.

NOBODY said it was impossible.

People doubted they could make it pay, but the US kept handing over money until it did.

Oh Starlink is still losing massive amounts of money. There’s absolutely no feasible way they’re profitable at this stage.

That said. I’m sure the US Department of Defense will effectively write SpaceX a blank check after seeing how quickly they switched to spread spectrum frequency hopping to prevent Russian jamming of comms in Ukraine.

> They have launched exactly zero of those "more powerful" satellites. And they are not launching at near the replacement rate for the final constellation.

You can't deny that they're going to launch the more powerful satellites while simultaneously relying on those more powerful satellites for your calculations of the size of the final constellation...

They're launching at much faster than the rate needed for replacement for the existing planned constellation for the version 1/1.5 satellites.

> Starlink will get slower and slower as they get more subscribers.

That assumes fixed capacity, which isn't at all the case.

No, it assumes that they will sell as many subscriptions as they can just barely support, because they desperately need the revenue to sustain the business model.

I was excited about Starlink until I learned more about it. Far too many issues. We need wire under ground or 4g. I lived in an area without many services (no cable, water, sewer) and had several satellite services which are more stable than Starlink and still found it inadequate.

What issues? I’ve had less issues than with other wired ISPs and satellite providers, personally. Not to mention any other satellite provider is going to be incredibly latency positive. Starlink has been a godsend. I guess just be careful of andecdata.

Nothing about the current situation is permanent. They're launching more and more satellites and it's only in the eastern half of the US that subscriber rates are growing faster than available capacity.

I think SpaceX would have to deliver downlink terminals at scale before third parties would be interested.

Definitely, and there are several other areas that need to be scaled up as well.

- Completing the gen 1 constellation, as well as getting gen 2 up.

- All of the ground stations

- Meeting the bandwidth requirements

However, there are potentially other ways SpaceX might be able indirectly benefit.

Another idea I've heard floating around is that Starlink can be used to link ground stations together. This might be very important in those especially hard to reach locations where laying fiber or establishing line of sight, etc might be too complicated/difficult/expensive. I've heard that this is not really a viable solution with gen 1s though.

I'm not sure how viable that would be, given that two of the reasons given for cutting them out would still stand:

- Not meeting the bandwidth requirements in practice.

- High initial hardware cost.

Starlink effectively being judged not able to meet the program requirements would surely rule them out to be used in connection with RDOF in any capacity?

I would imagine a $500 dish, plus installation (which can be done oneself) is much cheaper than digging trenches for fiber or copper lines, or hanging lines on utilities with obnoxious right of ways. For the cost of the dish to be cited as a blocker is mysterious, it's actually cheap compared to the prices quoted to truly rural folks to get fiber pulled to their remote locations.

It's a political move against Elon, just like the previous EV TAX incentive where the strings-attached involved having factories with union labor. Sure $500 is costly, but it's better than nothing.... that's the point.

> Sure $500 is costly, but it's better than nothing.

It's not better than nothing if you can't afford the $500.

Starlink's primary competition in those areas is likely to be 4G or 5G wireless using a fibre backhaul, not fibre to the network access terminal.

Yup, this is another idea I've heard floating around. In some situations where fiber backhaul or line-of-sight is not feasible, Starlink could link their ground stations to these hard to service locations extending the reach of existing backhaul.

Then the 4/5G can be set up using a mixture of fiber, line-of-sight or Starlink backhaul, depending on the situation and location.

Of course this is another way that Starlink might be able to win some of these contracts. They could set up (or work with a partner) mini 4/5G networks in these rural areas and also use a mixture of fiber and Starlink backhaul. After all, they still need fiber backhaul for their ground stations anyways.

Starlink customers end up not needing to buy the current, expensive hardware and just use relatively inexpensive 4/5G modems.

I've heard this isn't possible with the gen 1 satellites, so still a theory until the Starship program is off the ground.

I don't think this is correct. The FCC announcement notes that $5B of the original RDOF auction has been authorized already. It gave out $23M today in a separate announcement.

It’s weird… at that level of subsidy why can’t StarLink just give away the terminal for subsidized users? That’s over $1k per user, more than enough to pay for the dish.

I believe the issue is that they don’t have the bandwidth, not the dishes. They need more satellites to provide the speeds and amount of service they promised.

Exactly. Look at the starlink coverage map (https://www.starlink.com/map). Nearly the entire east of the US is "waitlist" aka at capacity. This is basically exactly what I predicted would happen - starlink is beyond awesome as a way to get internet but really suffers from lack of capacity "at scale".

Keep in mind too that each year average internet consumption rises so the existing sats can service fewer and fewer users.

I am personally surprised they have reached capacity so quickly - I thought they would have fit more users per sat on by aggressive traffic shaping/data caps.

Starlink will only be a niche service imo, FTTH/cellular will continue to roll out to more and more rural locations. It's a great niche service but the capacity really constrains it.

I also wonder how long SpaceX can afford all these falcon launches for starlink. I think the plan was spaceship would do huge deploys at a way lower cost per sat, which would increase capacity much quicker. But it seems spaceship is understandably delayed?

> starlink is beyond awesome as a way to get internet but really suffers from lack of capacity "at scale". Keep

Surely it pays for itself if they can sign up users as fast as they can scale? But if capital is the problem, maybe Starlink should be IPO'd the old fashioned way, to raise capital rather than cash out investors.

To IPO, they would need to release information that might make the whole enterprise seem less tenable. And, if it turns a profit they would be less able to siphon the money off to private projects. (No, not for a Mars colony. There will not be one.)

>(No, not for a Mars colony. There will not be one.)

Well, not with that attitude >:(

With or without, makes no difference.

Mars is wholly unsuitable as a place to colonize, and Starship would anyway be far from adequate to attempt it, by at least two orders of magnitude.

Starship might suffice to maintain a moon base, after the model of McMurdo Station, albeit radically more expensive. But that would be an end in itself, not any sort of staging point. There is noplace to go from there.

O'Neill cans make better sense. After aneutronic p-11B fusion is mastered in some future century, O'Neill cans out in the Kuiper Belt will be a place for humanity to expand. The inner solar system is a dump, Earth excepted.

Show the math.

Reader, meet exercise.

Writer, meet basic standards of making claims on the internet.

That reads like a claim.

Starlink seems to be as much real business as a way for SpaceX to raise money.

Winner!

The launches were delayed by COVID for supply chain issues, just like the antennas were. One rather surprising issue was that all the high-grade oxygen was being used in hospitals, with none left over for rockets.

https://www.businessinsider.com/liquid-oxygen-shortage-covid...

There's still tons of folks on 56k phoneline modems, <512k ADSL, unreliable&expensive cellular or expensive, slow and high latency VSAT connections in rural areas. In those places Starlink will dominate.

And the real global growth will start when the satellite-to-satellite links finally allow coverage away from SpaceX's teleports. So coverage in the middle of the seas and in the middle of Africa or South-America.

Yes but every year that goes by that market shrinks as new fibre gets rolled out and 4G/5G continues to grow in service. I don't know if it really is enough of a market to cover the cost of servicing thousands of small sats long term.

The global angle is interesting; but I don't think Africa will be as simple as people make out - there is a lot of corruption around telco contracts there and they won't just allow SpaceX to turn on all of Africa and bypass it all.

As they mass produce sats at ever increasing rates and at the switch to Starship and increase production rate of their termianls, their price per unit of capacity is gone go down much more. They have 3 vital parts all are dropping in price.

In addition, with laser links and improving routing software they still have a ways to go in optimizing the whole network.

The global shipping, flying and remote industrial use-cases are significant and pay a premium. In addition there are still a huge number of rural people in the world, Africa, Asia, Europe, South America. Then there are all kinds of mobile use-cases as well.

Starlink will also be used on things like rockets (like Starship) and all kind of military aircraft as well. Such usecases will also pay a premium.

Then there is global low latency backhaul as well.

I think if you really go threw it there is plenty of markets to service, even if you assume fiber is gone continue to roll out.

Exactly, the subsidy is meant as a "no household left behind" type program, so awarded companies need to prove they can cope with the best case scenario of 100% adoption, and StarLink probably wouldn't be able to cope with that many terminals in close proximity (typical rural community is sparsely populated, but not that sparsely).

I would imagine in the future they could narrow the scope of their bid just to really really really remote and unpopulated rural regions, where even a 100% adoption would fit comfortably in a satellite bandwidth, but then I don't know if such a smaller bid would be worth the hassle for them.

Per the article, the problem is that they can't cope with the users they have right now, not some theoretical future user count.

The FCC looked at data from Oopsla, and saw that Starlink is currently averaging 91Mbps Down / 9.5Mbps Up or something like this. Their bid was for 100 Down / 20 Up, so they are some way away from providing the level of service to their current users that they committed for the subsidized users.

But they have ~500k users now, with 20% of satellites deployed, and there are something like 600k subsidized households if I’m reading correctly. I think it’s safe to say that in 2024 they will be able to handle 600k extra users. (If they actually try and don’t oversubscribe!)

The 100/20 was observed early on IIRC so they have demonstrated it’s possible if they don’t oversubscribe. (Which is more than can be said about previous satellite internet services.) But yeah, they certainly aren’t currently in compliance. Given Musk’s general approach to regulatory bodies I can’t blame them for applying skepticism here.

As far as I understand from other comments, they will have the option to re-apply of they achieve these speeds sometime before 2028.

> I think it’s safe to say that in 2024 they will be able to handle 600k extra users. (If they actually try and don’t oversubscribe!)

We'll have to see. It's important to note that so far they've been able to use whatever launch capacity they want to increase the constellation, but by 2024 they will have to start using a good portion of their launch capacity to re-build it, since the satellites launched today will have fallen down by then.

If they manage to launch Starship by then (not sure if that's looking likely or not), maybe they will have extra capacity to be able to both replace satellites and expand the constellation.

Starship will blow the satellite count out of the water. Also, SpaceX gets to load up launches with payload space with more satellites.

What is apparent is that the other upstart satellite flotillas have zero chance of being competitive and getting off the ground, both figuratively and literally.

But at some point there were issues with getting components for the new gen of Dishys, they have less ports and flexibility than mine. Scaling dishys is still an issue along with the satellites, but it's not like growth has stalled.

This decision may minorly slow the growth of starlink, but not by much.

And if SpaceX does get Starship off the ground at anything close to their economic targets, fuggedaboutit.

Not sure why this comment was downvoted. I'm super pro-spacex, but you're spot on for all of these things.

That’s one objection, but another one in the FTC’s rejection as outlined in the article was the cost of terminal.

My point was “cost of terminal” seems like a weird one to be getting dinged on.

I agree capacity seems like a better thing to be focusing on. This seems like an own-goal from StarLink as they could have kept up their speeds by slowing growth and not enabling roaming. But maybe they are just prioritizing subscriber count over winning subsidy dollars.

Reply on news.ycombinator.com