you are missing some simple arithmetic here. yes, x times y can get arbitrarily big if x does (and y is non-negative), but then changes in y would just have even more of an impact. take x to be population and y to be emission per person and you should see your mistake.