Funny that you should ask that. In fact, the patent office rejects cold fusion patents. When Swartz tried to file some patents, he was rejected and sued; he ultimately lost (see Swartz, 232 F.3d 862, 56 USPQ2d 1703, (Fed. Cir. 2000)). The only surprise here is that the patent office managed to get this one right -- in my field (software), the patent office is totally fucked. If CF worked, it would be relatively easy to demonstrate: just generate a bunch of power with it on a small power input, in a way that others can replicate (even if you have to ship them your equipment).