Thunderbolt isn't competing with USB.
The two are apples and oranges: Thunderbolt 2 (which is the version in the announced, but not yet launched, new Mac Pro) is already at 20 Gbits/sec and is part video port, part data port.
However, the crucial difference is in the way they connect with devices at each end: Thunderbolt is, to all intents and purposes, a PCI bus on a rope. That's why each port requires a PCI lane. It interfaces with the computer (or peripheral) at a lower level than USB.
Thunderbolt, like Firewire, also does all the heavy lifting itself; your computer's CPU is therefore not bothered with running the protocol at all. USB, on the other hand, requires the CPU does some of the work. For consumers, the difference is academic, but for many professionals, anything that reduces valuable CPU capacity for their own projects is a big no-no. Why spend eye-wateing sums of cash on a high-end Xeon CPU and waste even a fraction of that power on USB processing overheads?
With USB, the CPU has to not only handle video data compression for my video editor application, but it also has to set aside some of its time to handle some of USB's duties during file transfers. This lengthens the time required for that video work, which is the work that actually *pays my bills*.
With Firewire or Thunderbolt, the CPU is *entirely* free to work on *my* stuff, not the connector's. My computer is therefore helping to make me more money per second. If I have to pay a one-off €29.99 for a cable to ensure that, I won't even blink at the cost as I know I'll make it back within a day or two.
So, no, Thunderbolt won't be a big consumer connection standard. Neither was Firewire. But the latter had its place in the high end prosumer and professional markets, and so will the former.
Last time I checked, Apple were rather fond of targeting both those markets. That's where the profits are.