They are Internet clients.
Internet clients use exponentially more data bandwidth than ordinary digital phones, which use thin streams of compressed data. Maybe several exponentials.
Operators are thinking of expanding their networks into homes and offices to handle the extra load.
The Android clients shown in Barcelona aren't much. What they mainly prove is that the specification can be built and deployed quickly.
While the Apple iPhone roll-out has gone at a predictable pace, with one vendor delivering specified numbers to a handful of networks, the Android roll-out will be far more helter-skelter.
That's because the Android is designed to work anywhere, first on any GSM network and then, with a little tinkering, anywhere else.
This makes it hard for operators to predict where and how demand will come from.
While Internet routing and fiber trunks have always been scaled well ahead of demand, mobile networks will have to route a ton more data on-and-off the fiber for the new clients. Equipment will have to be bought without an assurance of a quick return.
Yet thanks to the iPhone network operators have no choice. In the U.S. Verizon and Sprint are seeing their best customers jump en masse to AT&T, because of the iPhone. They have to compete.
While the Google Android specification may (or may not) allow effective competition on the handset end, Verizon must guess that it will, anticipate that demand, and start investing now.
It's a gamble, in some ways bigger than the spectrum auction was.