Interesting post. My background is I've been in the IT field over 10 years.. but I've been supporting ADSL in the states for 2 different internet providers (my current position is with a ILEC and I've been withworking for them for 4+ years).

I can tell you that all the grid sounds like to me is nothing more than a big fiber backbone. Which many companies already employ using SONet for countrywide transmission. The devil is in the details though and in order for that transmission medium to reach your house it would 'have to use existing high speed lines'. So in the end, it's still bottlenecked on the front end.

In short, I fail to see how structurally they are achieving anything which is not already being done.

Regarding ADSL in the UK only acheiving 6mbps; the standard of g.dmt/G.lite is only written for 8mbps. Realistically even with a sub 4000 foot loop, you're looking at BEST around 6-7mbps. Frankly, if they're advertising 10mbps on standard ADSL, they are lying to you. Frankly, the standard and technology in no way can support it. I know as part of our infrastructure upgrade we had to upgrade our systems to ADSL2+ to offer a true '10mbps' service here.

Now if they are using 10mbps, they would need a rack of ADSL2+ cards on the DSLAM (yay technical terms) which could support it, along with appropriate bandwidth on the back end to handle the transmission of that data to their regional aggregator.

It's possible they are running ADSL 2+ equipment with insufficient bandwidth. Which means upping their speeds will have no impact until they upgrade their pipes. In fact, it should actually make things worse due to more packet collisions and poorer resource management.

Anyways...

The next big thing for traditional copper is pair bonding. This will increase the speed a bit more (roughly 45+mbps) after this evolution of adsl 2+. There's also the "RIM" chip which I read about somewhere which is supposedly in development to offer 200+mbps over a standard POTS network.

/end geek rant.