Niks, I think you're a little confused about diffraction gratings. Beam splitters will split the beam up into weaker beams, yes, but so do diffraction gratings. A 100mW laser through a diffraction grating will not create many 100mW beams, that would be even better than perpetual motion! Unfortunately you can't pull power out of thin air so easily. Using a series of beamsplitters will make each individual beam weaker, but no worse than a diffraction grating. I think using a diffraction grating would lead to a worse overall loss of power (as some beams would have to be blocked out, unlike with beamsplitters), though it would be a far simpler solution. Either way, if you have X beams from a Y power laser, each beam will be no stronger than Y/X, period, and in all likelihood will be a good amount less than Y/X (EDIT: Well, I guess that technically only applies if all beams are equal power, but still, you can't end up with more enegry out than you put in).
I think you're also underestimating power. 100mW laser becoming undetectable after beamsplitters? Even 5mW is an awesome amount of power and we normally take it for granted. A photodetector would easily pick up a few microwatts or less. The only concern would be lesser visibility to an audience (in this case, even 100mW is weak), but a diffraction grating far from solves the problem. Unfortunately, the only solution in that case is good old fashion money. The more you want, the more you'll be paying. To make a visible harp for outdoor audiences like I believe the Jarre guy uses, even the most expensive Hercules laser would fall far short.