Granted, TD-LTE's sales pitch is not all that different from its ancestors - i.e. making use of unpaired spectrum to boost capacity in urban environments where FDD macro networks get overloaded. What is different this time around is a bigger ecosystem of vendors developing it - admittedly for just a single market at the moment, but also the biggest single mobile market in the world.
The other key difference is that TDD has always been primarily a data play. But from 2001 up to 2008, 3G cellcos were still primarily in the voice business, and FDD allowed them to continue milking that cash cow. That worked fine when 3G data usage was still mostly ringtones, wallpapers and other walled-garden content.
Then the iPhone happened. Smartphones got smarter and data usage skyrocketed so high that E1 backhaul links became the new bottlenecks. If ABI Research is to be believed, by 2014 mobile users will be transmitting a total of 1.6 exabytes a month (compared to 1.3 exabytes for all of last year).
Hence all the interest in LTE, as well as related technological tricks to offload data traffic and maximize RAN capacity like spectrum refarming in the 900- and 1800-MHz bands and femtocells. TD-LTE is another tool in the toolbox, and by the time we start hitting monthly exabyte levels in five years, its predecessor in China will have been put through the ringer enough to qualify as "seasoned" if not "mature".
Of course, all that depends on a ton of factors over the next five years. Still, TDD is a lot closer to realizing its potential than it was at the start of the decade.
If nothing else, TD-LTE may have the novel distinction of being the quietest evolution the cellular world has yet seen. That will depend on how much progress Qualcomm and other chipset vendors make with dual-mode FDD/TDD chipsets, but once devices are capable of roaming seamlessly between both, TD-LTE may be the first RAN acronym that won't need to be marketed to end-users who don't give a toss what it's called anyway.
Which will make a nice change.