In this case we are not really talking about a free consumer product, though.
I would be surprised if Toronto's municipality would grant this development without having some guarantees of it being carried out, save for paying severe penalties.
Google Wi-Fi mesh network in Mountain View hasn't worked in a long time. Toronto may well get similar service decline.
Also, Starbucks Google Wi-Fi rarely uses Google Fibre, IME, and instead uses commercial providers like AT&T and Comcast. Plus, Starbucks hasn't deployed much SGWF in Silicon Valley so it's still basically vaporware. (There's a Starbucks in Santa Clara with non-Google Wi-Fi as slow as dial-up.) They'd be wise to cancel the Google partnership and go with an outsourcer whom has access to a couple of big providers to maximize cheapest/fastest/coverage and handles all the deployment, support and security.
Maybe. AT&T started deploying FTTN to major cities like Houston and Chicago in 2005-2006 (four years before Fiber). Their gigabit buildout was based on that original plan (which dates to SBC). Comcast has been on the DOCSIS upgrade treadmill for years. In the 7 years before Fiber they went from 8 to 100 upgrading to DOCSIS 2.0 then DOCSIS 3.0. Gigabit is another 10x jump over a similar time frame, based on DOCSIS 3.1.
There was also Verizon FIOS, which started in 2005, expanded to 5 million customers, then stopped expanding much after 2010. It apparently cost Verizon more than they'd expected.
Actually, I hope not. Nearly every building in this neighbourhood has fibre to the home, there are three different carriers with fibre into my condo. Most of the rest of the city has copper. While it gets a bunch of press, I hope this is more focused on applications and IoT than pipe.
80% sounds quite high but I wouldn't be surprised if the number is around there. Google are big enough to take risks and try things that don't work out. There are 95 official Google products in the Wikipedia 'discontinued products' listing - https://en.wikipedia.org/wiki/List_of_Google_products#Discon..., but that's not very accurate in that it doesn't mention things have failed but are limping along, and it includes things that were discontinued but live on in other products (eg Wave merged with Docs).
It's not entirely unreasonable to suggest Google have 20 successful products for their ~95 discontinued ones though.. which would work out as 80% 'failed'.
Why is that significant? How many projects has GE or IBM discontinued? Google is at the peak of its profitability today, so whatever is being done to products seems to be working out.
