There wouldn't be a serious problem raising $140b from financial markets if doing gigabit-to-the-home were profitable. We basically did an equivalent level of build for CATV 1980-1995, and for cellular 1985-now (still upgrading and adding sites). The question is financial return and how long you can depreciate the asset.
I'm curious what the breakdown is between rural areas (where, presumably, construction costs per mile are far cheaper), suburban areas, and rural areas (where you can hit a lot of people per building, but costs to shut down a road are higher, and where there are few options for routing.)
It's depressing how hard it is to get 1Gbps into random buildings even in Silicon Valley. Technology has made 100M fairly feasible in most offices, at least (you can do channel-bonding on an old T1 circuit to get 100M, even without fiber!), for <$3k/mo. It's still hit or miss if you're on-network for municipal fiber (in PA), the alternative fiber providers (like IP Networks in SF), etc.
Given the actual average cost for bandwidth is less than a dime per gigabyte, working out to less than $50 per month per household, I'd say it's extremely profitable. [1]
The only reason no one is pursuing it is because the incumbent providers have lobbied tirelessly to create the notion that no one can afford to deploy new fiber, which is clearly false.
If we were to convince regulators to push hard on telcos and recover the rural broadband tax, we'd have probably nearly half a trillion dollars to pay for it. [2]
What's the information value of this article? Hmm, it's "bad news for those of us who would love to see Google (GOOG) expand its fiber network out to the entire country".
Did Google ever mention they're shooting for 100% of the country in the near term? They did not. Are they likely to follow traditional ISP model and launch market by market? Yeah, from what we've seen that seems like a reasonable assumption. Is Google likely to borrow $140,000,000,000 to blanket the entire country? Past experience says probably not.
Australia is in the midst of doing pretty much exactly this. Called the NBN. One of the main political parties who are in their second term started it, and hopefully they get a third term so they can finish it. The other Political Party has said that if they get in they will stop the rollout. It's been budgeted at approx $40 billion+ to do.
Hopefully it will complete as at the moment we dont have any data guarantee's for the last mile. In fact if you can get a dial tone on the 100 year old copper, thats good enough for the old telco(which was privatised about 10 - 15 years ago).
My opinion (and its different in australia due to our different views on private vs public investment, market size, country size, population etc) is that most large infrastructure just cant/wont be built by private investment, because it just isnt profitable. Diving into it I would say that most power plants, telecommunication, water infrastructure, roads etc are mostly funded by the government even if that is through tax breaks etc. So if the citizens want infrastructure the govt needs to pony up the cash.
It also hasnt been lost on a lot of people here, that we sold off our government owned telco 10-15 years ago, only to create a new one not to long after. Ahh The cycles of govt.
I don't see how that's "cold water" - yes, covering the entire continental USA with anything is very expensive. They don't have to buy the whole thing at once now, do they?
I don't think Google wants to build this at all. They just want to put some pressure on Comcast, TW and the likes to offer higher speeds and without any caps. Google wants to bypass Windows and go straight to the cloud, but one needs high speed internet.
Cable companies are deep in debt for a reason. I just don't see Google hiring 80,000 people are whatever it takes and hope it just works financially.
They'll need to make money in Kansas or something like that. Then they'll have more money to roll out and expand in other areas. It may take years, unless google sparked competition that others ISPs start to upgrade their fiber too.
Additionally it's not clear that Google sees the KC experiment as anything other than an experiment. Money is not the only rubric by which Google Fiber's success should be measured.
Just having the data on what running Google Fiber took would be super useful and interesting (and in the event that it does fail, I hope that info makes it out into the wild).
I wonder if Goldman's estimate includes the cost of permits or not. One thing that bothers me about deploying fiber in the easiest city first is that it's not representative of the political environment in 90% of the US.
Given the response to Google's call for selection just from municipalities, I'd say the political environment is extremely favorable. Several cities have shown a willingness to rush through new ordinances to make it possible to issue permits quickly.
It's only the pressure from incumbent cable and telco operators that has held back municipal fiber deployment across most of the country. Once that pressure is pushed aside, cities are willing to move very quickly.
I'm curious what the breakdown is between rural areas (where, presumably, construction costs per mile are far cheaper), suburban areas, and rural areas (where you can hit a lot of people per building, but costs to shut down a road are higher, and where there are few options for routing.)
It's depressing how hard it is to get 1Gbps into random buildings even in Silicon Valley. Technology has made 100M fairly feasible in most offices, at least (you can do channel-bonding on an old T1 circuit to get 100M, even without fiber!), for <$3k/mo. It's still hit or miss if you're on-network for municipal fiber (in PA), the alternative fiber providers (like IP Networks in SF), etc.