I think the point is that if you move towards rent-seeking vs. value creation, people with a critical mass just hoover up capital and get richer. Even if the economy contracts, you can still take more chips off the table.
We have historical examples of this as populations start bumping against local ceilings that illustrate that this scenario is very possible. The Roman aristocracy was living it up, while the plebeians in many cases were subsisting on bread and circuses.
So unless there is a relatively low limit for total humans on this planet, we won't reach it before we become a spacefaring race, at which point the limit will be the carrying capacity of the Universe.
But even in that case, you can only have at most polynomial growth (human habitat spherically expanding with speed of light) instead of an exponential one.
The mechanism of r getting bigger than g is not an increase in r, but a decrease in g. The increase of the global economy is dominated not by increasing technological efficiency, but by population growth. There is an upper limit of the number of people we can fit on this rock (whether 9 billion or 100 billion, we will hit a limit at some point). Once we bump into that ceiling, growth will slow dramatically, and those who have already managed to create a sizable stash, or who have the talent to regularly beat the market, can accumulate wealth to no end.
In short, as I understood Piketty's book, the fact that r might not increase if we put the right limits on housing development, doesn't change the overall hypothesis made by Piketty.