Minecraft Forums

Actually, they had two costs it appears, one to pass the home ($900 in '06) and one to hookup the home ($650 in '06).
According to Gizmodo, they were paying $750 per home and $600 to connect.
http://gizmodo.com/5503428/verizons-fios-rollout-drawing-to-a-close
Since that was two years ago, and costs seem to have decreased by a good sized amount since '10, we can assume it actually costs $600 per home to pass it, and $500 to connect it.
Now all of a sudden my estimate is seeming somewhat low, mostly because I forgot to factor in the two different costs.
Let's rerun those calculations. I'll even drop the cost to $1000 combined.
$1000 pass/connect, $75 router/modem, $25 tech visit, $300 offset, $800 left over.
$70/mo - $20/mo (bandwidth), $15/mo (combined network/service costs). $35/mo profit, $800 to pay off, over 22 months before they're profitable. Ouch!
Rerunning calculations with fiber ISP at 100mbps....
$1000 pass/connect, $40 router/modem, $25 tech visit, $0 offset, $1065 left over.
$80/mo - $3/mo (bandwidth), $15/mo (combined network/service costs), round up to $20/mo. $60/mo profit left, over 17 months to payoff/profit.
Rerunning calculations for EPB's pricing (100mbps, $140/mo, no setup fee)...
$1000 pass/connect, $40 router/modem, $25 tech visit, $0 offset, $1065 left over.
$140/mo - $3/mo (bandwidth), $15/mo (combined network/service costs), round up to $20/mo. $120/mo profit left, over 8 months to payoff/profit.
One last re-run for EPB's 1gbps pricing... (1gbps, $350/mo, no setup fee).
$350/mo - $20/mo (bandwidth), $15/mo (combined network/service costs). $315/mo profit left, over 3 months to payoff/profit. Not bad.
Things to note: Pricing on tech, modem and combined costs is not 100% accurate. Bandwidth pricing and oversell ratio is hard to determine, but these costs should be accurate.

Verizon has disclosed that it costs (on average) roughly $1000 per home for the fiber infrastructure/buildout. Let's assume costs have declined to $700 since then (quite a reasonable estimate). Based on consumer pricing for equipment (with a massive cost decrease factored in), I'd assume their modem/router/switch combo costs Google $75 (might be a little low). We'll also assume that a tech visit to hook up the fiber is at least $25 out of Google's pockets (it'd actually be higher, but we'll ignore this). So, $200 of that install fee is left. After applying it to their infrastructure costs, their remaining expenses to cover would be around $500. Now, this ignored the whole one-year contract stuff... But I'll ignore it for now.
So, we start out with Google being at -$500, and getting $70/month. They are now obligated to supply a 1Gbps line to the consumer. Let's say that each household will require... 200mbps of backbone capacity (not bandwidth, capacity). So we're looking at 50 households per 10Gbps line back at the PoP. That's actually a little dangerous for the future, as having only 20% capacity online and ready at any given time isn't a good idea. Eh, whatever.
With that in mind... We'll assume each household uses 20mbps of bandwidth, so that's nearly 6.5TB set aside per household. It seems like a good amount for the present, but nowhere near enough in the future.
So, 10Gbps line + 1Gbps bandwidth = 50 households ($3500/month). I wonder how much they'd pay in transit costs? We'll assume Google is getting a very good price, so $1/mbps (which is absurdly cheap). $1000 in bandwidth costs, $2500/month of profit left over from that 50 house chunk ($70 in, $20 in bandwidth costs, $50 left over). Now, let's look at the money Google needs to cover other stuff. I'll list them right here!
• Support
• Marketing
• Payment processing
• Sending bills
• Dealing with fraud
We can easily add $10/month per subscriber for all of that.
Now, we're at $40 left over for Google, $500 to be paid off.. And wait, we're forgetting something! That $700/user fee? Sure, it covers most of the infrastructure. It's missing other costs though. Power for the ISP's equipment and network, equipment (pumping a gigabit to everyone is VERY expensive), and of course more (like all those technicians you need who monitor and maintain the network 24/7, and that's just for the network — we're not even talking about install/service techs yet). Oh yeah, and you can't forget legal fees, taxes, replacing stuff as it wears out or becomes outdated and handling copyright issues.
Hey, let's look what the ISP's central office is going to contain. A very expensive bundle of fiber links, a very largeand expensive bundle of customer fiber links, millions of dollars worth of switches, routers and network gear, a impressive and obscenely expensive UPS/generator system, another impressive/expensive HVAC system, a expensive and large (and secure) building, some content/mirror/DNS/billing/internal/backup servers, and much, much more.
Calculating how much of each subscriber's bill is going to be chewed up by this is not a easy task, but I'll say it's safe to assume a minimum of $20.
So Google is making maybe $20/month profit after all this. Oh ****, we forgot about that $500 of negative income after every install. Well, at least in 25 years they will have paid for the install costs, and some of the other costs will have came down.
Now, there are a number of variables here. Google could be paying more for bandwidth, the router/modem could cost more money, the $20 I removed for infrastructure could be shrank to $10...
Hell, let's just combine and shrink the $10 and $20 amounts to $15, and ignore the rest.
Okay, Google is making $35/month after all this, but it'll still take 14 years to pay off the money they spent hooking up the customer.
You know what? Let's re-run these calculations (with a different ISP), but assume that:
• This new ISP only supplies a modem
• This new ISP is only providing 100mbps (let's say for $100/month)
• This new ISP is planning on profiting from this
• This new ISP is going to oversell at the same 5:1 ratio
• This new ISP does not charge an install fee
$40 modem, $700 ISP hook-up investment, zero install fee kickback, $25 tech hookup (by the way, this number is extremely low, if we wanted to be more realistic than it'd be much higher). 5:1 means that a 10Gbps/1Gbps backbone = 500 customers, 20mbps speed/2mbps (650GB) bandwidth reserved per customer. ISP pays $1.50 per 1mbps, so $3 in bandwidth per customer.
$100 in per month, $15 fees to pay, $3 in bandwidth cost, hell we'll assume they this ISP is paying more and round it up to $20/month. So they're raking in $80/month, making back that $755 initial investment in only 9.5 months! After that, they can enjoy $80/month of profit.
There are a few holes in my calculations. Modem and modem/router/wifi combo pricing is most likely too low, initial hookup costs for the technician are too low, bandwidth costs and overselling ratio is uncertain, and the amounts of bandwidth being reserved are weird. However, it's still a pretty damn good estimate in my opinion.