One of the problems in venture capital today is a fundamental impedance mismatch: fund sizes remain large, while capital requirements for many Internet/software deals are shrinking.

You don’t need much money anymore for many software ideas: the software stack is free, servers can be rented for $50-$100/month, and there’s cheap labor offshore. There are a lot of ideas that can be vetted for $100k to $1m.

At the same time, venture funds have grown and stayed big, driven in part by VC compensation. As I wrote in an earlier post: venture isn’t generating great returns these days, pushing VCs to make their money on fees. The larger the fund, the larger the fees.

The mismatch happens when you do the math: for a $200m fund with 4 partners, each partner needs to invest $50m. If each partner does 1-2 new deals/year, and the fund is committed over 3 years, then each investment has to be a $8-$16m commitment. (That doesn’t mean that Series A needs to be $8m, but it means that the total invested is in that range).

You can see where it is hard for many firms to do $1m investments — it’s just too small. And some of the most interesting stuff is happening “down there”!