Your numbers are way wrong, even if we discredit another commenter's point about city vs metropolitan areas. The top FIVE cities in the US total over 19 million people. If you instead use the top FIVE metropolitan areas you get just shy of 70 million. That's 20% of the US population living in or directly around the top five cities. Furthermore, there are over 300 cities in the US with populations over 100k. I'm too lazy to do the math to add it up, but I'd venture a guess than 50% of the US population lives in a metropolitan area of over 100k people.

(Edit: fixed 25 for 19, accidentally added a city twice)

Many small/medium American "cities" are municipalities that are directly attached to (or even surrounded by) a larger urban area. It's basically a quirk of how local government is structured rather than anything that has to do with population or infrastructure density. For example, it's unlikely that many residents of Beverly Hills (<50K residents) think of themselves as living in a small town outside of Los Angeles.

I'm not sure where you got your numbers. The 10 largest cities in the US have a population of 26M (2015 estimates, Wikipedia). Generally population estimates are given for the metropolitan area. The 10 largest have a population of 73 million. There are 53 MSAs with a population over 1M with perhaps a total population of 150M. (I was too lazy to do all the addition.) The US has a large area, but it is primarily urban.

> The largest 10 cities are home to 16M Americans, the 10 largest cities in the UK are host to 12M Brits. The 100 largest cities in the US are home to 59M Americans, the 100 largest cities in the UK are home to 34M Brits. The US is 5 times as populous but there is only 30% difference between the top 10 cities in the US vs UK and 50% in top 100.

This is a nonsensical statistic to use, irrespective of whether the numbers of right or wrong. Imagine taking the UK and copy-pasting it 10 times. Make the new deca-UK a single country, let's call it 10UK.

Obviously, 10UK has exactly the same proportion of people living in "major cities" as the original UK. Obviously, the infrastructure serving the 10 copies of each major city is equally affordable for a country with 10x of everything (or even more affordable, due to more economies of scale).

But the 10 largest cities of 10UK are the 10 copies of London, and thus the proportion of people living in the 10 largest cities of 10UK is equal to the proportion of people living in the single largest city in the UK. Clearly, this is the wrong number to use to estimate anything related to the problem at hand.

This is one of those rare situations where avoiding the wrong conclusion requires zero knowledge of the world; cognitive ability on its own is sufficient. In other words, you are demonstrably stupid.