Hacker News Reader

Comment by rayiner | Hacker News Reader

Some back of the napkin math. $140 million for 40,000 households in Boulder = $3,500 per household. If 50% of people subscribe (typical for municipal broadband), that's $7,000 per subscribing household, or higher than Charter's market cap per subscriber.

A quarter of the population of the city is below the poverty line. If the intent is to offer them subsidized broadband, let's assume we get no capital recovery from those households. That brings the cost to $4,666 per household, or $9,300 per subscribing household.

This is high, but not out of the ballpark. Verizon spent about $3,500 on average for each paying subscriber for FiOS. Chattanooga's system was about $5,500 per household.

View on HN

$1.1m/year payback (20 years)

$1,100,000/year, 40,000 households, so $27.50/year mean income per household to service the debt.

If only 50% of households (you said people) subscribe, then $55/year mean income per household to service the debt. If, again, as you state, no capital contribution from households below the poverty line, then, using your math (which, btw, double counts these), we're talking $110/year, or $9.16/sub/month to service the debt.

Here in Austin, there are three providers of 1gbps/1gbps FTTH internet, and all are priced just under $60/mo.

I think someone can probably figure out how to build a similar service in Boulder, paying the City of Boulder $10/mo per subscriber household, and still make fat rolls of cash.

That $1.1 million is just for the $15 million initial phase, not the $140 million full buildout.

Exactly. This makes so much sense for the city. This type of calculation doesn’t even include a fuzzy bump in technological growth for a city that’s connected with cheap fiber.

The monthly cost of fiber in Austin has gone up a bit since you last checked. Google and Grande charge $70/month. AT&T is up to $80.

Repaid over twenty years, as I understand it: "A 20-year debt structure would require annual payments of $1.1 million from the city, according to staff estimates. That could be supplemented with money from the city's general fund and, perhaps, revenue generated by early customers of the fiber network." Though there is some math discrepancy since that would not cover the $140 million. Nonetheless, as a Boulderite, I am happy to finally see this happening.

That's about 10 million per year in debt payments for the full $140 million investment. If you assume that 50% of households subscribe, and the city subsidizes rates so no capital is recovered from folks below the poverty line, that's about $686/year per subscribing household. Which is crazy high--the non-subsidized ARPU will have to be $140 to break even. (Which is also not out of line--Verizon will barely break even on FiOS at $110 ARPU, and that deployment was cheaper and had no subsidy component.)

Comcast charges $160 a month for 1000 Mbps internet and $110 a month for 400 Mbps. The $100 per month you quote isn't terrible when compared to those numbers.

I wish people would also list the upload speeds, which for doing backups, sending grandma 4k video links of the kids, etc is miserable on the 20-40Mbit, that the cable companies are providing to go with most of their "gigabit" packages.

The idea that you need to use facebook's/etcs spyware just to share a couple photos with your friends goes against the entire premise of the internet's PtP architecture.

More generally, does a media companies ownership of the infrastructure create a conflict of interest when it comes to providing a symetric communications path so you can stream your kids video's instead of having to involve a 3rd party to "cache" them for you?

Unless Comcast is doing a fiber installation to your home, it's not comparable. They are probably just raising the caps they have on their existing conditions, but the neighborhood is still sharing the same amount of bandwidth. It's also not symmetrical up and down.

If I was an employee of Comcast, when someone asked what I did I would be so ashamed I would rather say "i am a prostitute that uses the money for drugs".

$100/mo for FTTH vs. $69/mo for 250MB/s Comcast?

I'll take the FTTH and leave Comcast and Century Link behind, thank you very much.

Would be even better if the City charged you for the fiber link and then let you select your backend provider. E.g., the way consumer choice works in the electric space. Let the city run/own the cable, and let the back-end carriers compete to provide the backbone and value-added services. Give every house a basic connection that lets them interact with city and state services only, for free.

> $100/mo for FTTH vs. $69/mo for 250MB/s Comcast? I'll take the FTTH and leave Comcast and Century Link behind, thank you very much.

It's $100/mo only if you hit 50% uptake rate. If your uptake rate is only, say, 33%, because most people don't care about gigabit and would rather save the $30/month, your per-household cost shoots up (as does what you have to charge to recover the investment).

Maybe you're willing to pay more. Most people probably aren't. (FiOS uptake rates are under 40% even though its always been much faster than competing cable services.) And the brutal economics of fiber really punish you for that. If your uptake rate goes from 50% to 33%, your per-subscriber costs have to go up to compensate (which puts even more downward pressure on your uptake rate).

FIOS includes TV. The real issue is Verizon is not that much better than cable companies.

I'm not sure I understand your first point? TV is a major part of the economics of fiber deployment. Chattanooga's fiber system also includes TV, so does Google Fiber. As to your second point--have you ever had FiOS? Verizon's wireline division is run like old-school AT&T--FiOS is very high quality.

I have used FIOS for the last 5 years. For a while they where throttling Netflix just like the Cable companies, so better sure but they also wanted to kill net neutrality just like the cable companies.

As to TV, you spread the network costs over both internet AND cable TV and should include a portion of TV/equipment rental revenue when calculating payback periods. Cord cutters are significant, but many people still want cable TV channels and rent DVR's.

I agree you need to include TV revenue--that's why I threw in Verizon's $110 ARPU (which includes TV), and suggested that Boulder's more expensive system would have to hit an even higher number like $140 (including TV).

The average Comcast customer paid something like $150/mo in 2017.

Replace that with a $100/mo. FTTH plan, plus Netflix and you're still saving money. And many people have Netflix + Comcast already.

I image that $150/mo average include cable and landline? I spent between $35-65/mo at my last 3 apartments (CA, KS, MO) for ~100-200mbps through Comcast for internet only.

An extra 50 a month is definitely worth not ever dealing with Comcast support or their shitty network.

So $200 million for the entire build out including debt costs? Your debt calculation assumes a Muni-bond rate of 4% which seems high to me. AAA muni bonds are more like 2.8%.

I'm impressed by your math skills, but I used to live in Boulder and I'm shocked to learn that 25% live below the poverty line. Is that really true? I suppose it could be the fact that many are college students, but still. That seems high.

Those trailer parks in North Boulder are very well hidden. Many don’t even realize they’re there. Some of the nicest trailers I’ve ever seen, FWIW.

There's also the hippy co-ops like the radish.

There's also a ton of kids waiting fir their trust fund to, vest I guess?, and don't really care that they aren't making any money at the moment.

Actually there are trailer parks in, I think, 4 different places in Boulder.

Most cities try to hide trailer parks as much as possible. 99% Invisible recently did an episode on mobile homes that was pretty interesting.

I don't think that's true. Those parks aren't big and they're still expensive because the property itself is expensive.

Most of the poverty number is due to all the college students - there are 33k of them in a town of 108k.

A rising tide doesn’t necessarily lift all ships.

The advantage of having this subsidized by Boulder is that the city will capture some of the increased productivity due higher speed internet in the form of taxes, both for individuals who are able to earn more money and companies that benefit from better internet infrastructure.

I've also read about studies that show that access to the internet improves health outcomes for poor people, so Boulder may also benefit from fewer emergency room visits and fewer sick days from those low-income households.

Which makes for a pretty compelling reason as to why Charter hasn't done it themselves.

That, or maybe $140M is a gross over estimate, or is referring to more than what’s discussed in the article? I’m tending toward that; it can’t possibly cost anywhere near $10k per subscriber to build out fiber normally, even in small cities, can it? Seems like if it was normally that expensive, there wouldn’t be fiber anywhere. What’s so different in Boulder? I’d speculate that CenturyLink paid less than $1k to run fiber to my house, even if I include my share of their neighborhood & citywide costs.

Costs in Boulder are quite a bit higher than other cities because we have very little to no topsoil in areas, you hit bedrock (which is often granite) immediately, and you need to drill or blast that out for most underground construction.

At least that was the line Google fed us when we got rejected for their fiber, who knows.

That could explain it. My fiber comes in from an above-ground wire strung on poles. They’re proposing to bury the lines, at high cost?

poles are uncommon in residential areas here, everything is buried. and digging is expensive.

It seems dumb that poles are uncommon in an area with little topsoil over a granite substrate? Why so fancy? Incidentally this could mean that the trailer parks are cheaper to serve than swankier neighborhoods; there are lots of poles in trailer parks.

It's Boulder, houses are getting close to California levels. The cheapest starter homes are around $700k.

there are at least 3 trailer parks in boulder, maybe more. i wonder how prices in them have changed over the years? :)

A single wide is about $300k, but there's a five year waiting list.

FWIW, the market appears to have cooled down a bit this spring. It remains to be seen whether that's a blip or a trend.

winter storms and strong winds. i've never lost power in the boulder area, except when a substation failed, regardless of the weather. when i lived in areas with overhead lines, power loss was a regular event, occurring during winds that would be considered mild in boulder.

Regularly trimming trees around the power line seems less expensive than boring through granite? The frequency of storm-related outages is a good proxy for how much the power company is spending on vegetation removal.

It's quite a battle as we also get gale force winds coming off of the mountains so often that it's not news, just a normal Tuesday.

> Regularly trimming trees around the power line seems less expensive than boring through granite?

That's not going to do anything to stop ice from forming on the lines, which is a far bigger problem in the winter than tree falls.

The front range gets 100mph+ chinook winds, and late spring storms dump feet of heavy snow leading to what we affectionately call "treemageddon", during which town will be filled with the thunder of trees crashing down for hours and hours. It's not an ideal environment for lines on poles.

at least in newer developments, my understanding is most underground channels (think sewers) generally provide some facility for running new lines/equipment in them, to reduce the number of times that city streets need to be dug up to lay x,y or z.

I heard on the grapevine that the real reason was that city council tried to hold the new campus hostage as a bargaining chip (nice campus you have there, it'd be a shame if the property was re-assessed and greatly increased your tax burden) but Google was having none of it and pulled the plug then and there.

I have not studied the whole Boulder area in detail, but as a PhD Geologist and having visited friends in Boulder that have added a basement to their house, I'd say you might have been fed some misinformation. Maybe little topsoil but lots of glacial till. They might be right about digging being expense though. Glacial till can be full of huge boulders that are hard to trench.

I live on glacial till, can confirm it's a huge pain to dig through, just in my own backyard. Drilling holes for anchors, I spend more time prying fist sized rocks than drilling. Which is incredibly difficult in a four foot deep, ten inch wide hole.

There are a couple areas where there is tons of topsoil sure, like in South Boulder I think many houses have basements, but as soon as you get out west or north you'll see all granite. Being a geologist I'm guessing you know the wide range of formations and different geologic timeframes we have across the span of like two miles.

> it can’t possibly cost anywhere near $10k per subscriber to build out fiber normally, even in small cities, can it?

I don't know, because I don't possess any expertise in the matter. What I do have is a history of watching public works projects budgets balloon completely out of control. It seems far more likely, given history, that this estimate is low.

> It seems far more likely, given history, that this estimate is low.

Which history do you mean? A lot of people have fiber to their house, that’s part of history.

If the cost has ballooned out of control, why? Costs don’t have to do that, and they haven’t always done that historically.

Fort Collins is in the midst of this right now, they proposed borrowing "up to 150 million" to roll their municipal fiber out. The cost break down is here, page 5.

https://www.fcgov.com/broadband/pdf/Broadband_Executive_Summ...

I got a survey from Larimer County about having municipal internet. Of course, it doesn’t seem like it will make it to my place outside of Estes Park. I’m stuck with a 5mbps WISP :/.

I work for an ISP, typical cost for just labor + materials for fiber is $10k per mi., labor being the majority of the cost. I live in the midwest and fiber techs start getting paid around $25 per hour and it can take a crew of 2-3 guys for a few hours per household which if you have in-house guys isn't as much but if you're contracting it out can inflate cost 2-3x. That's if you don't have to make the run down the street and live close to a node.

That quarter below the poverty line is probably a little misleading in the way it's calculated.

Students there are all living off parent money so their disposable income is much higher than hard data would suggest.

Does all this math include building out the call center, technician and customer support, field cars, and on going maintenance? Building it is one thing, maintaining it; not just the cable but the hubs and data center; plus staffing with all be costs to be accounted for. Throw in do they need to subsidize or provide computers for the low income residents as well.

It isn't just lack of availability of internet that keeps low income out its the lack of devices to make use of the internet and rudimentary training.

we don't really care how much sewer or electrical connection cost. It is just a must-have utility connection, even though we know that it is possible, at least for some time, to live without it. The municipal fiber is the same. Not everybody looks at it this way today - like not everybody 300+ years ago saw why municipal sewer is necessary.

these are great stats.

I live in Chattanooga.

I know 3 residential users of gigabit fiber. (I'm not 1. 100 is perfect).

Don't get me wrong. The "gig" here has made a major impact. My company is based here and the change in volume/quality of startups from 7 years ago to now is staggering.

But... I'm not sure why Boulder needs this.

What is the expected lifetime of this kind of fiber infrastructure? Shouldn't the city also accumulate funds for eventual replacement?

It depends. The actual fiber is probably good for decades. But a lot of the active infrastructure requires upgrades in a 10 year or so cycle. FiOS started out in 2005 with BPON. Then they upgraded to GPON. Now they’re working on NGPON-2. This requires upgrading ONTs, (very expensive) OLTs, switches, not to mention adding capacity out from the COs to handle the additional traffic.

OLTs aren't that expensive, especially since you can potentially split the costs of each port over 128 subscribers. Obviously the total dollar costs add up when you put thousands of subscribers on one OLT, but the cost per subscriber is quite low.

NGPON-2 isn't being deployed at scale yet, so prices are still to come down, but GPON OLTs are bulk items. You can buy a GPON OLT for a few grand which ends up being a few bucks per subscriber.

Verizon upgraded to GPON in 2008, when prices were a lot higher. Back then, GPON line cards were in the $10-20k range. (And they're upgrading to NG-PON2 starting this year.) Also, their typical split is 1:16 or 1:32. So you're talking about $600+ per subscriber (maybe pushing a grand once you factor in cost new ONTs and cost of labor to install new ONTs). That's a couple of thousand dollars per subscriber over the amortization period (at least if you're hoping to keep up with the Joneses).

Your numbers are way off.

In 2006, Verizon's cost to pass a property was $850 and trending towards $700 in 2010. This includes not only the outside plant, but also the OLTs. The majority of the cost is in the outside plant, so the GPON OLTs are only a fraction of this.

Verizon deploys at scale and does not pay list prices. Furthermore Verizon would not be deploying NG-PON2 if it wasn't profitable and would definitely not be doing it if it was cutting into their margins. Therefore there is no reason to assume their NG-PON2 costs will be higher than their GPON costs.

Well, it also will likely be a huge boost to the economy, so it will pay itself back in other ways too.

Reply on news.ycombinator.com