clock menu more-arrow no yes mobile

Filed under:

Index Ventures' Mike Volpi on Enterprise Investments

The Cisco exec-turned-venture-capitalist breaks down his investments in Hadoop, software-defined networks and search.

Index Ventures

Index Ventures has been busy over the last few months. At least five companies in its portfolio have gone public, including Zendesk (last month) and Arista Networks (last week). Others in the family appear to be headed in the same direction, including Dropbox and Pure Storage.

Earlier this week, Index announced the closing of its latest fund, its seventh. It seemed like a good idea to catch up with Mike Volpi, a general partner at Index and former chief strategy officer at Cisco. He has led several of Index’s most recent high-profile investments and sits on the boards of Hortonworks, Pure Storage, Sonos and Zuora, among others. He spoke to Re/code by phone earlier this week.

Re/code: Let’s start with the news: You just closed a new fund. What’s this one going to focus on?

Volpi: The fund is the new Index Ventures VII, and it’s 400 million Euro or about $550 million. It’s essentially a carbon copy of what we did last time on Index Ventures VI. In terms of focus, it’s also similar to previous funds. We’ve done a lot of consumer-Internet-related things, with a concentration on mobile, marketplaces and networks. In the recent past we’ve been a little more active in financial services. And on the Enterprise side, we’ll do applications, cloud and cloud infrastructure, and Big Data. We’ll occasionally invest in low-level infrastructure like storage, and then we’ll also sometimes invest in security as we feel that’s an interesting and controversial space.

With Europe, do you tend to focus on certain countries or is it pan-European?

In principle, we’re pan-European plus Israel. In the rear-view mirror, you end up with a few concentrations, and those are around London, around Stockholm, around Berlin and around Israel. Obviously there are offshoots. We’ve made investments in Portugal and France and Spain, but really those four geographic hubs are where we have tended to invest.

You’ve also been a pretty strong investor in some enterprise-facing companies, and you’ve had some recent exits with the IPOs of Zendesk and Arista Networks. What are the specific themes you see playing out there?

The overarching theme has been disruptions in cloud and software-as-a-service, which are sort of cousins, and mobility. 20 years ago, when you built a new enterprise technology, you had to go after the banks because they were always the early adopters. Now, those early adopters might be someone like Twitter or Box or Zynga. And then you see a bank like J.P. Morgan or a government agency like the National Security Agency take advantage of them, too, because of the cost or efficiency gains they offer. It’s a target-rich environment because there are lots of places where you can find early adopters. Broadly speaking, that’s a very good thing.

You’re a significant investor in Hortonworks, which is one of two or three companies leading the charge to get companies using Hadoop to tackle their big data problems. Obviously, you can’t talk about Hadoop without mentioning the huge investment that Intel made in Cloudera, just as Hortonworks raised its own nine-figure funding round. What is it specifically about Hadoop that makes for such a huge opportunity?

Historically, much of the data that is now going into Hadoop clusters was simply not collected. There was too much of it, it was too expensive to collect and store. With the advent of Hadoop, we now collect a lot more data than we did before. It can be customer clicks, it can be the wireless company keeping information on where your phone has been and what it has done, it can be medical information. We used to see it in an instant and then throw it away. Now we can actually keep it. And that information — if you look at the trajectory, we’ll keep twice as much data in the next 10 years as we kept in the last 50.

Historically, the information we kept used to sit in a relational database like Oracle. And the reason that Oracle became Oracle is that the relational database is at the center of almost everything we do today. Over the next decade, what you keep in a relational database will be the minority, and the majority of it will be in Hadoop. And that means that everything else you do with your data will be designed with Hadoop at the center of it all. Hadoop becomes the underpinning of the computing architecture of the future.

So what opportunity do you think Intel sees?

The large majority of the chips it sells in 10 years will be associated with Hadoop clusters. And that is why so many companies care about Hadoop ecosystems. Even though Cloudera and Hortonworks are both open-source companies, and there have been a lot of differing opinions on whether you can really build a large profitable company around open source technology, we certainly believe that you can.

The size of Intel’s investment certainly raised a lot of eyebrows. Relative to the importance of Hadoop, it’s probably not a lot. On the other hand, it [raises] the question: If Intel were going to spend that much, why not just buy the company? The question that was never answered was around some of the terms. Does Intel get a right of first refusal or did they get certain terms to facilitate a future acquisitions? At Hortonworks, we took investments from financial companies. And so we feel like we are putting the company on a trajectory to a public offering.

You’ve invested in Pure Storage, which I’ve written about, and also Big Switch Networks. That’s a software-defined networking startup. SDN is supposedly another big disruption that’s coming to networking infrastructure, yet not much seems to have happened yet. How is Big Switch doing?

It’s doing great, actually. It took a little while for the SDN concept to come around. I spent a lot of years in the networking business. The lesson I learned there is that the hype cycle comes way in advance of the reality. In 1989 everyone talked about DSL and then it took about five years for it to roll out. Similarly, we talked a lot about SDN a couple of years ago when in fact very few customers were ready to roll it out. In the process, Nicira was acquired for a lot of money before it had much revenue. Now I think the technology and market readiness has come a long way. Big Switch went through a period of relative quiet. But now it has quite a nice list of customers. Revenue at this point in the year is double what was expected, and it has a number of product announcements in the pipeline for the summer. We all got caught up in the SDN hype about 18 months ago, and then we had to go heads-down because the market wasn’t buying. Now there are real customers. When you ask customers to change how they build their networks, it takes time.

You’ve also invested in a company I’ve been hearing about lately, but which I don’t know much about, called ElasticSearch. What’s that one about?

When I talk about our portfolio and mention ElasticSearch, we get a lot of people saying that we’re cool. Developers really like it. It’s an open-source product and when we invested it was clocking maybe 150,000 downloads a month, and now it’s above 700,000 downloads a month. And here a download means a developer is installing it in their infrastructure. We think of search as just one use case like Google, but there are so many other use cases for search. If you need to search through a system’s log files to find errors, you need search. If you’re Booking.com and you want to help a customer find the ideal flight, you need search. If you’re an investment bank and you need to see where the anomalous trades were last week, you need search. It’s sort of this omnipresent use case where anywhere you need some kind of search, ElasticSearch is kind of running the tables on it.

This article originally appeared on Recode.net.

Sign up for the newsletter Sign up for Vox Recommends

Get curated picks of the best Vox journalism to read, watch, and listen to every week, from our editors.