<Nitpick mode on.> It seems the average IQ of the US is 98, not 100. But even if the mean would be 100, and <100 would be "dumb", it does not follow that 50% of the population is <100.

Actually, since IQ is bounded on the low side and not bounded on the other side, it is actually likely that if the average is 100, more than 50% are below average. But that is not guaranteed. You could have only one dumb person with everyone else >100.

IQ is well known to be normally distributed. One property of the normal distribution is that median = mean, so it follows that less of the population would have an IQ below 100.

I’m no expert, but isn’t it a normal distribution by definition?

Still nitpicking: Since IQ<0 do not exist, it cannot be a true normal distribution. It is true that IQ distributions over large groups resemble normal distributions in their core, i.e. close to the median.