The VC panel at Structure 09 was quite interesting, as was the case last year. There was uniform reticence about investing much in capital expensive infrastructure: one VC said that when someone comes in with a pitch on how they will be more cost effective than Amazon they just laugh. This evolved into a deeper discussion on the problem that it poses for investing in valuable infrastructure management software and systems as well. There is growing appreciation that it makes sense to learn in large Cloud environments and use that learning to design products for the broader market (this is Microsoft's System Center strategy for example) which makes the larger Cloud providers very interesting early customers to serve. However a very large fly in the ointment comes from the Amazon business model: Bezos is a retailer, and like all successful retailers, is very good at cost minimization. As he tells his investors, the good part of having a low profit margin is the large barrier to entry it creates for smaller competitors given Amazon's experience and economies of scale. As an expected consequence, Amazon, like WalMart, is very good at squeezing suppliers, including software suppliers. Google proudly pays no software license fees on their 2MM servers, and Amazon clearly pays as little as possible. So if you hoping to make good money selling innovative stuff to Amazon and learn from the experience (or any other large Cloud provider) think again.

Comments