Having worked at the largest-small ISP in the US (>250k subs). Very ambitious, and I think it can be done, however your biggest problem is in the over-build.

Having worked at the largest-small ISP in the US (>250k subs). Very ambitious, and I think it can be done, however your biggest problem is in the over-build. What that means is you will have to engineer and deploy your own HFC (hybrid fiber coaxial) network first. If your original intent was to lease space on an existing provider's network, and potentially buy some frequency on their coax you're in very steep uphill battle. Even if a provider did happen to let you do this you're beholden to them at that point and they, bluntly, have you by the nuts if you start to threaten them. Meanwhile overbuilding where there already is infrastructure isn't very efficient as you have to plan and engineer (if you have no experience with this think civil + RF + data engineering collaboration plus legal hoops for right-of-way and licensure into actually laying fiber and coax). Add to that along the way you'll piss people off for digging in their yards. Sure, you can sell channels a la carte, however say you purchase 5 channels - and you have 50 customers. Say every person of the 50 only chooses one to subscribe to, now you only have a ratio of payment for 10 customers per channel. This will likely put you in the red right there since each channel has a pretty steep licensing fee for distribution. Once you get to scale - this could work, but it would almost be easier for you to go in and buy out a small customer base instead of starting from scratch. On top of all these considerations realize if you're offering Internet and/or phone you need CMTS and phone switching systems plus the engineering clout to run them. If you plan on spanning multiple cities then you have to deal with leasing or running your own fiber and then building out transport back-haul (Infinera is an awesome company for DTN platforms BTW). Long story short - you need a lot of smart people to help you design and build this out, and if you've never worked with any ISP engineering verticals you've got a ton to learn. Keep in mind this equipment is pricey and the only way to make a good profit is to own your networks end to end (we were turning up 1 & 10 Gb circuits like hot cakes after we installed a new 1Tb system and since we owned all the infrastructure we were looking at tacking on another Tb right behind it to keep pace with demand - but keep in mind all the Tier 1 connectivity you need to support this as well). Good luck!