Not sure what your point is, but that's exactly how surveys work : take a sample of size n (10000) and do some statistical inference based on the observed statistic, i.e. either reject or accept a hypothesis based on the sample.
10000 seems statistically significant
Way better than the n=30 they teach in stats 101 anyways