JVP - Jerusalem Venture Partners

By Bob Brown

Erel Margalit, founder and managing partner of Jerusalem Venture Partners (www.jvpvc.com), takes a broad look at the network industry, both in terms of what the companies he invests in offer and where they operate. As the name of the venture firm implies, many of the companies JVP invests in have their roots in Israel or have offices there. Margalit himself used to be director of business development for the city of Jerusalem. He spoke last week with News Editor Bob Brown about where JVP is putting its money these days, and why.

Q. Given Israel's reputation as a hotbed for network security technology, I'm surprised JVP doesn't have more security investments. Why not, when this market is supposedly so hot?

A. What you always worry about with security plays is that they're too narrow. We recently funded a company called Cyber-Ark (http://www.cyberark.com) the company is based in Cambridge, Mass., but was founded by a group of Israeli engineers and has offices in Israel], and that's the first time we've funded a security company in years. The reason we went with Cyber-Ark is because it addresses the idea that the main security-related worry for the enterprise is being able to share information with the external world and with different locations across the enterprise. Cyber-Ark came up with a very interesting product that allows connectivity in a secure fashion. Instead of you opening your own gateway to your enterprise, you take a relevant cluster of files and databases and put them in a virtual safe run by Cyber-Ark.

A lot of security companies have been focusing too much on the issue of security, whereas you need to focus on the functional thing that the customer wants to do, solve that and make security one of the things you're addressing.

Q. What other network technology developments are coming out of Israel?

A. There's work on next-generation semiconductors for storage devices. We're also seeing with one of our companies, called Dune (http://www.dunenetworks.com), the need by the big optical systems companies to outsource a lot of the innards of their routers or network devices to people who can make them cheaper and modularly expandable. I still think some of the most interesting, disruptive technologies are coming out of there on the software side, largely the software behind enterprise systems products.

Q. In light of the renewed fighting in the Middle East, how much of a business-as-usual situation is it for these companies?

A. People forget that when the venture industry began in Israel we were in the middle of the first intifada. The way most Israeli companies are set up these days, the headquarters are abroad, so these companies are really sticking to their business plans. Believe it or not, as far as Israelis are concerned, there's relative calm there right now. It's actually a good time to make investments in Israel because a lot of the international funds that were clouding the market are much more comfortable following a fund that's active there than taking the deal away from us.

Q. What's the mood surrounding the network industry in countries outside the U.S. where you've been of late?

A. A lot of the clues of what's going to happen in the U.S. can be found outside it. Coming from China, Singapore, Taiwan and Europe, I see that the moods are not equivalent around the world. There's a lot of demand for new technologies in the Asian markets, for example, where there's an amazing amount of 802.11 wireless infrastructure going in. VCs that can help their companies approach more than one market can sometimes help validate a business model with enterprises or carriers elsewhere, so that they come to the U.S. with a little more maturity.

Q. Looking back toward the U.S., among your investments is Cogent, seemingly one of the few CLECs left standing. Is it really worth staying in this market?

A. We are seeing a whole generation of CLECs wiped out, but we are seeing perhaps one, two or three next-generation players that will be able to compete with some of the incumbents. Cogent (http://www.cogentco.com/home.html) is a little different. The company is very well financed and what we're trying to solve is the problem of getting fiber to the buildings. Once you have the fiber all the way to the customers' buildings, the customers will come. The problem is that a lot of these CLECs were carriers' carriers, and at the last mile they just used the copper or DSL or wireless or what-have-you. They weren't able to bring the broadband, the Ethernet all the way to the enterprise. There have been too many plays in which these next-generation carriers have provided part of the route rather than owning the customer from point to point. Cogent is going building by building, targeting the 3,000 largest in America to start with. One thing we did with Cogent was take advantage of the downturn in the market by gathering assets very cheaply from struggling carriers. One was Allied Riser, which had a pretty amazing amount of risers into some of the largest buildings in the U.S. And they had some customers. We also bought PSINet's assets in the U.S. out of bankruptcy and that was a network that was worth a few billion dollars just a while ago and we got it (http://www.nwfusion.com/news/2002/0301congent.html) for [$10 million]. With that we got the second best peering agreements in the U.S. behind those of UUNET/WorldCom, so that gave us an additional 18,000 buildings that could be relevant to our network. One other thing we're doing with Cogent is partnering with vendors of different services that run on broadband networks and target vertical markets like legal and financial services. Cogent should have very significant revenue this year and hopefully be EBITDA [earnings before interest, taxes, depreciation and amortization]-positive next year, along with a lot of money in the bank.

Q. So Cogent can really give the incumbents a run?

A. It's absurd to think that nobody is going to compete with the incumbents on data provisioning to the enterprise, because that's traditionally where they've been the slowest. They're not innovative or flexible. If you have a company with a network that is really just built for data, you're running the company on reasonable costs, you're able to put more and more customers on your network and you're turning your wide-area network to speeds of local-area networks. That's what the industry has been looking for. Of course, if you look at the institutional market, there's still a lot of skepticism, so you need to build critical mass and show that you are profitable. And it still may take a year or two for the market to appreciate you. You also need to distance yourself from the disasters that have occurred in the industry so that people can accept that there could be a fresh start.

Q. Despite the fact that so many CLECs failed, has their activity served as a wake-up call to the incumbents?

A. The likes of AT&T and Verizon are talking about new strategies, but if you're competing with the incumbents and have a good strategy for data services, then you're in OK shape. The incumbents are still gravitating to a lot of the voice part of their business, and they still have a lot to do before they invest in the Ethernet rings that are going to be alongside the SONET rings.

Q. You've also been focused on the carrier markets from the optical components side. Are there any indications from the components vendors that demand for carrier gear (and possibly new services based on that gear) is picking up?

A. For a lot of people, the optical components sector has become unfashionable. What's happening now though is we're taking the stars of performance - the companies that came out with 40G bit/sec lasers for long-haul network gear, for example - and diverting their energy and using the new materials and processes to deliver carriers equipment that not only boasts better performance, but also lower costs. The Nortels and JDS Uniphases of the world are partnering with these new technology makers to create super components, using the technologies for the metro, not just the long haul, which has been brutally silent. A lot of the demand has been on the metro side, for building 2.5G bit/sec and 10G bit/sec networks. We're still a ways from this technology making it to the enterprise network.

Q. With Juniper's recently announced plan (http://www.nwfusion.com/edge/news/2002/0520ju2.html) to buy Unisphere, some observers say we could finally be seeing the start of an industry shakeout. Do you expect merger and acquisition activity to pick up anytime soon?

A. Cisco bought some companies a few weeks ago, too. I think we're going to be seeing more of the large players going back to buying companies and technologies because the prices are still relatively attractive. Once the large players are getting a bit more visibility on what they actually need, they're going to go and get it from the outside world.