It is absurd because if BT or AT&T wanted to upgrade certain areas to the new network to improve service to their consumers and gain additional profits out of it, they wouldn't choose cities where there already is a competitor with an equivalent fast network.
For example, if AT&T is convinced they are capable of providing a speedy 1 Gbps network, when Google announced they were going to Kansas City last year, AT&T could have decided to go to Austin and improve their service in the area.
However, this is not how it's playing out at all. Instead of choosing another densely populated area (only 2 cities in the US have this type of network, so there's plenty to choose from), they select (or signal) they will go only to where the competition is, and that is pretty absurd IMHO (not from a gaming theory perspective though).
Either way, Google's strategy in this case is flawless, and regardless of what other network providers do, it signals to the consumers that the technology is available and at reasonable costs. Therefore, it will start creating the push for network upgrades in cities where Google never intended to go in the first place at all, which is what this is all about. Having a fast and reliable network available to as many users as possible serves Google's interests (incidentally the consumers needs too), and that's what they are pushing for.
Regarding AT&T, they are signaling being willing to bring to the ground any business models relying on such upgrades, which as relevant as it could be to Comcast or Verizon, it's meaningless to Google as these investments are literally sunk money to defend their traditional business model.
In the same way they already defended in the past for other entrenched markets which could undermine their advertising business model, such as: Android (mobile market, avoiding iOS powerful position) or Chrome (internet browsers and default search provider settings).