I'd like to back up a bit, and ask, why?
There are many amplifiers you can buy off the shelf, and they'll generally work pretty well. There are different tradeoffs of cost, power output, fidelity, power consumption, etc. What is it that you expect to do to improve on what's already available? Or are you just planning to try and learn something, without necessarily improving on an existing design? Are you wanting to learn how the design process works, or would you be happy just to assemble something based on a schematic someone else developed?
You say you want to drive a 100W speaker. Just one, or do you want stereo output? How much power do you want to put into that speaker (or those speakers)?
For a low power, low voltage audio amp, you could start with the LM386 data sheet, or search the web for LM386-based circuits. If you want more power, both Sparkfun and Jameco offer kits with more power available. If you want something even bigger and better, search for "carver amplifier schematics" for some examples of audiophile amplifiers. These are just a few ideas -- audio amps are so common, there are lots of schematics available.
One note of warning: Because of the economics of the electronic parts supply chain, if the product you want to end up with is something that's already available off-the-shelf, then it's really hard to build something yourself that's better and cheaper. If you have some unusual requirements that aren't met by off-the-shelf equipment, then building it yourself is the only way to go, of course. And sometimes it's fun to build things yourself, despite the fact that it may be more expensive and/or worse performing.
Anyway, that's why I'm asking "why?". We need to know your motivation in order to offer constructive suggestions. Do you want top audio quality for the best price? A commercial unit may be your best bet. Do you want to learn how op-amps work, and how to design passive and active filters? Working through the relevant exercises in Horowitz and Hill's "Art of Electronics" might be a starting point, though it will take quite a bit of time to accumulate the knowledge and experience to design your own high performance amplifier. Do you want to copy an existing design, without necessarily understanding it thoroughly? Maybe someone can offer some suggestions for kits or schematics.