Age is not the only thing that could raise an individual's risk factor to 10%. There is no reason why individuals not in their 80s might, for some other reason, be in a category with higher mortality risk. Honestly, we don't know that much about this disease yet - but we do know that the outcomes for different people can be very diverse.
One problem with age-based stats like this is that it gives the impression that the way to think about risk is along the lines of russian roulette: when someone gets infected, the great actuary in the sky rolls a D100, consults the tables, and if their roll is good, they don't die.
But what if the die was already rolled for some individuals? What if, say, one in 100 people carry a genetic risk factor that means that if they get COVID, they have a much higher risk of dying - regardless of their age? That hypothesis is also compatible with the statistical data we are looking at.
You could also consider, what if some people have particular predisposition to become a superspreader and infect many more people?
If the population contains individuals for whom their individual outcome risk, or their potential impact on other people, is significantly above the average, then by reducing the spread of the disease we reduce the number of such people who ever get infected. You can prevent the D100 having to be rolled for as many people, including people who need to roll much higher to survive.