Yep, it does take up twice the space .. but no, it is not something that should concern you. Java will take care of it and your PC has more than enough space that you could easily declare everything to be a double.
For 32-bit computing, a float requires ('takes up') 32-bits of space. A double takes up 64 bits (twice as much .. double ..
For 64 bit computing, these numbers are doubled.
Yes, doubled. You'll find that almost everything in computer science gets doubled or halved A LOT. I'll leave it to you to figure out why.
Why don't we declare nothing but doubles? There are a couple of reasons.
One: just because we have the space does not mean that we need to use it. It is sloppy to assign resources 'just because we can'.
Two: there are many applications where this small amount of space can make a difference. Not every codes for PCs ... there are also phones, PDAs, etc. ... and these require that programmers are more conscious of their space decisions.
Three: a double takes up two 'natural spaces' in your PC. A 'natural space' (commonly called a 'word') is 32-bits for a 32-bit computer and 64-bits for a 64-bit computer (etc). It is the amount of data that fits into a single 'slot' in memory. There are many advantages to keeping your data in line with the size of the 'slots' in your PC, but they are beyond the scope of this discussion. Let's just be Martha Stewart and say that using floats where possible is A Good Thing.