Sramana Mitra: Reversing the question now, if you look at 2017, what have you seen? What trends have you seen in your deal flow? How many companies do you see in a year for example and how many do you invest in? What has been the 2017 deal flow nuggets of trends?
Rajeev Madhavan: In 2017, my average was eight companies a week. That’s 400 companies a year. Out of 400, we invest in six. It’s almost like one in hundred. One of the things that I believe very much in is giving them the feedback even if we’re not the investor. I have some empathy with them as I have gone through what they’re going through.
Sramana Mitra: What are the trends in there?
Rajeev Madhavan: Last year was the year where AI is in everything, whether they have data or not. It’s a common pitch. AI-as-a-Service is never going to work. It’s going to be the applications that AI uses and changes on its head. Things that could be done by human beings, they can be automated. There are two challenges to it. Number one is, can you get the data and can you own the data?
For example, companies came to me in semiconductor saying, “I can do yield optimization using artificial intelligence.” They say they’re going to get the data. I’ve been in that space. You won’t get that. You really need the ability to have the data that you need for doing AI. Sometimes that means striking a partnership and then the actual end business application that gives you more value is what’s going to give you the benefit of using AI.
Out of our 13 companies, nine of them use AI. I would call seven of their AI applications as more predictive learning rather than real, deep AI. Two of them are real, deep AI. Last year was the year of artificial intelligence. There were 11 chip companies that we saw to all the way in different applications making claims that AI is going to be the be-all and end-all solution for all. We think that AI will play a role in an application is being written or done but the end value is going to be in the vertical use of that AI in an application. That’s one of the theses that we had during the last year.
Series A was getting a little tighter while seed is getting a lot easier. Silicon Valley has had a lot of excess with the angels. Seed was getting very easy. Series A was getting tougher. It’s almost like there’s a middle glut of money. Then at tail-end where Series D and E are, there’s a lot of money. Seed has a lot of money. Part of it is due to people’s extreme goals in terms of valuation. This whole unicorn impact drove them to think that you’ve got to increase your valuation arbitrarily. At the end of it, valuation in the early stage is not going to be the differentiation. This is from experience.
Sramana Mitra: Ridiculous valuations are not a favor because eventually, to get a success for everybody, you’re going to need to get an exit. An exit is not going to be based on your last round of valuation. It’s going to be based on your traction.
Rajeev Madhavan: It’s actually very negative. Then you go for the next round and if it’s going down, it’s actually going to hurt you tremendously. Even in IPO, there are rights that you’ve given to the last investor. You’re in a terrible position if you’re going with that model. You really need to think through that and not worry about early-stage valuation.
You have to make sure that it’s a win-win for you, for all the employees, and for all the investors. It is absolutely the kiss of death because you really want to give your entrepreneurs an opportunity, especially to make their first few big contributions valuable. If you do not do that, you’re not going to be able to recruit in Silicon Valley.
If you want to recruit great talent, you better have talent that is better than you as a founder. That’s a given. I think I was one of the dumbest guys in some sense in Magma. There was a period when everybody in the team had a Ph.D. except me. I really loved the fact that that was the case. Hire people who are smarter than you.
Sramana Mitra: There are different kinds of smarts. PFor pople who are doing electronic design automation software, having solid electrical engineering credentials is very helpful.
Rajeev Madhavan: That’s true in every one of the areas.
Sramana Mitra: I mean that as a proxy.
Rajeev Madhavan: Yes. There are founders who come in thinking that we’re going to pay them like Google is paying. You need to understand that’s not what’s going to get a person to come in. You got to get them excited about the stock. The moment you’ve done that, everyone will work towards the same goal of success in the company.