They don't have the same kind of board power they used to, but they're still very influential. They are often more experienced than founders, and also, being older, more authoritative. And much of their advice is good. Especially the best investors, some of whom are really smart. So founders rightly start to rely on their advice.
The catch is that founders' and investors' interests are not always aligned, and when they aren't you can't trust investors' advice quite as much. But the two types of cases aren't sharply differentiated. Which means to avoid being misled by investors you have to be able to judge precisely how misaligned your interests are in each situation, and discount their advice by exactly that amount. It's a subtle problem.
So the leverage is either official (maybe they have a board seat) or (more often) social - founders feel either that they SHOULD listen to their investors because they're older and more experienced or that they have to for various reasons. Makes sense.
One thing I always forget about being a founder is that you're constantly reaching places where you have absolutely no experience - and yet you're still in charge. The Collison's for example are younger than me, and I'm pretty young, yet they continue to run a company that is growing at an extraordinary rate. Must be pretty interesting.
I thought he meant "what leverage do the founders have?". Not sure though....
However, assuming that he did mean "what leverage do the founders have?", I would say that they are the boots on the ground and can very much scuttle anything that the investor wants them to do. The moment founder's interests are not aligned with the startup, it is doomed. I guess an investor can help them succeed but can't stop them from ruining it.
Disclaimer: pure speculation, I have no experience of being on either side of the table.
The catch is that founders' and investors' interests are not always aligned, and when they aren't you can't trust investors' advice quite as much. But the two types of cases aren't sharply differentiated. Which means to avoid being misled by investors you have to be able to judge precisely how misaligned your interests are in each situation, and discount their advice by exactly that amount. It's a subtle problem.