Wouldn't that make one giant 1.625% milk?

That's a good point. It seems to be an infinite series that should converge on one value of leche. If we solve for that, we should be able to make the ultimate torta de leche, which, paradoxically, is made with only one supreme leche.

What about condensed milk or powdered milk? Do they count as distinct leches as well?

I didn't want to show my hand, but off the record, I've experimented with Norwegian beaver milk and Mongolian camel milk to achieve in vitro samples of sesenta y tres leches. Admittedly, it's highly unstable and there have been casualties. Such is the price of progress.