I call FUD. 64-bit is only "what people are after" because of marketing. Nothing more or less. I mean, think about it, what really is the point of 64-bit?
64-bit integer maths isn't really a genuine requirement, and on the rare occasions it is needed the impact of performing 64-bit integer maths on a 32-bit CPU is not too immense. As for 64-bit floating-point maths, most ARM chips have come with this built-in for many years.
Then there's 64-bit addressing, which in reality is a myth, since no CPUs actually support 64-bit addressing. Nobody needs to access 16EiB of RAM, or will need to for several decades to come. I believe that x86-64 chips currently top out at 48-bit addressing, which is 256TiB. 32-bit ARM chips top out at 4GiB, which admittedly is starting to feel a little cramped and is arguably inadequate, but the Cortex-A15 introduced 40-bit addressing (1TiB) which addresses this concern.
The reality of "64-bit" for x86, and the performance advantages it has brought over IA32, has been that it's addressed deficiencies of Intel's old IA32 architecture. The main improvement derives from the addition of 8 new general purpose registers, bringing x86-64's tally to 16. ARM chips have always had 16 general purpose registers.
I'd argue that ARM have already designed cores that are capable of playing in the laptop space. Cortex-A15 MPCore seems up to the job to me.
If you're still not sold on my arguments that you don't really need 64-bit, ARMv8 was announced last November which is a 64-bit ARM instruction set. Applied Micro's X-Gene CPU is based on this.
Besides all of this, given that their business is designing cores rather than manufacturing it's not really down to ARM to push into the laptop space. It's down to their licensees to put ARM cores into laptop CPUs, and to manufacture them using processes that will allow those chips to run at clock speeds competitive with Intel and AMDs CPUs.