When investing in early stage companies, you put in less money and the upside is much greater. If this were the full picture then obviously early stage ventures would be a better deal. Of course, many more early stage companies go under than late stage companies. This extra risk is supposed to balance out the greater reward. But does it?
To borrow Kedrosky's terminology, VC investing is really a combination of two separate skills: picking and poking. Picking is choosing which startups to invest in, and poking is helping the startup along to further its chances of success.
The vast majority of people investing in early stage ventures are friends, family, and fools. Because these people aren't very good at either picking or poking, their risk level is very high. This in turn drives up the cost of capital.
You can reduce your risk by becoming better at picking. You can also reduce your risk by becoming better at poking. And if you can become better at both, you can reduce your risk a huge amount. Because there are so few investors who are good at both, these people are barely even taken into account when determining the cost of capital. Which is why if you are good at both, it's an arbitrage opportunity for the investor.
For late stage ventures, there is much more money on the table which means much less amateur investment. Because you don't have as many dumb people pumping up the cost of capital, there isn't really any arbitrage opportunity.
Incidentally, PG has reduced his risk even further by creating news.YC so he can get to know the founders a little bit before investing. Theoretically, if a founder knows that PG has less risk because of this, the founder should be able to use this as leverage to get a better valuation. However, this doesn't work in reality because the shortage of PGs creates a market inefficiency (in favor of PG). However, I suspect that anyone who can successfully clone him will have a good arbitrage opportunity.
Well Alex makes an interesting point, that you would need more skill to get a big return on the little hands, but I still think that if you had a large sum of capital and you were especially good at the poking, that small investments would pay off exceptionally well.
To borrow Kedrosky's terminology, VC investing is really a combination of two separate skills: picking and poking. Picking is choosing which startups to invest in, and poking is helping the startup along to further its chances of success.
The vast majority of people investing in early stage ventures are friends, family, and fools. Because these people aren't very good at either picking or poking, their risk level is very high. This in turn drives up the cost of capital.
You can reduce your risk by becoming better at picking. You can also reduce your risk by becoming better at poking. And if you can become better at both, you can reduce your risk a huge amount. Because there are so few investors who are good at both, these people are barely even taken into account when determining the cost of capital. Which is why if you are good at both, it's an arbitrage opportunity for the investor.
For late stage ventures, there is much more money on the table which means much less amateur investment. Because you don't have as many dumb people pumping up the cost of capital, there isn't really any arbitrage opportunity.
Incidentally, PG has reduced his risk even further by creating news.YC so he can get to know the founders a little bit before investing. Theoretically, if a founder knows that PG has less risk because of this, the founder should be able to use this as leverage to get a better valuation. However, this doesn't work in reality because the shortage of PGs creates a market inefficiency (in favor of PG). However, I suspect that anyone who can successfully clone him will have a good arbitrage opportunity.