|
Most people think of Tom Jermoluk as the guy who used to run @Home Network, in the years up to and months just following its merger with Excite. But a longer look at Jermoluk's professional past shows other links to broadband, including his top role at Silicon Graphics in the early '90s, helping Time Warner Cable develop its Full Service Network. And there's a decidedly technical bent to Jermoluk: In 1988, he won accolades for a paper titled "High Performance Polygon Rendering." Jermoluk, who goes by "TJ," has been described over the years as everything from "broadband's warrior" to "party boy" and "boyish, irascible." Broadband Week Technology Analyst Leslie Ellis caught up with Jermoluk recently to look into his new role in the venture capital world, at Kleiner Perkins Caufield & Byers. An edited transcript follows.
BBW: It's been 9 months or so since you uncollared yourself from the CEO life at Excite@Home. How's the VC world treating you?
TJ: It's great. I'm happy. Which is a lot more than I could've said a year ago.
BBW: What? You don't miss your friends in cable TV?
Jermoluk: Some more than others. (laughs)
BBW: And now you're working with the most visible venture firm in Silicon Valley, smack dab in the middle the whole dot-com financial mess. Were you affected?
TJ: We stuck to the knitting. That means we didn't open an East Coast office, or a European office. We did what we know how to do, and we're in a pretty good position today, because of that.
With the downturn of the market, and all these companies going under--now, new companies looking for funding are realizing that all these VCs formed in the last year or two don't have any more to give. That makes them more desirous of a real VC. So the pool is even bigger now than it was a year ago, if you can believe that.
BBW: And what about your decision processes? Have they changed with the downturn?
TJ: It would be foolish not to recognize that the downstream funding requirements have changed. Radically. For example, a company with a business plan that stipulates a tremendous amount of capitalization--$100 or $500 million, like an access provider--that's harder money to find.
BBW: Thinking back, @Home, raised, what, a billion dollars? That was arguably an access provider.
TJ: And can you do that in today's environment? I don't think so. That's why all these DSL providers are going under, and access providers are going under. It's not just lower valuations, there's no valuation that people will invest in.
The same is true for b-to-c companies. Somebody who says, "I need $400 million for advertising and warehouses," now hears, "sorry, don't know where you're going to find it." So, yes, we're saying, show me the business model, show me that you really do get gross margin and profit. Then show me the financing plan, and how you'll achieve it.
Kleiner has four rules about startups that they taught me. Four risks to work to remove: Technical risk, management risk, market risk, and financing risk. Technical risk is, can you make it work? That doesn't apply to selling pet food on the Web, but it does apply to free-space lasers. Management risk is, can you put together the right talent to make it work? If you're thinking big, can you pull together big? Three is the market. Do people want it? Does the dog want the dog food? Four is, is it fundable. Does it have a business model you believe in, and can sell to enough other people?
A lot of VCs didn't go through the risk sets. They figured the capital markets were in infinite supply.
BBW: You mentioned the talent pool. You used to worry about that--attracting and retaining good people. Has the downturn eased that?
TJ: It's gotten better. It's not like there's any engineers standing on the street corner (laughs). Again, they're going for the quality environments. That helps firms like KP. We're viewed as having the high quality track record.
BBW: What of all the bizarre stories about people coming out of college with expectations for huge salaries, food brought in for every meal, car services, Foosball tables in the break room?
TJ: I had one the other day. We really liked him. He really liked the company. We made an offer. He says, I was expecting more like this (higher number). I said, you're a year late. It's like me trying to sell my @Home stock. (laughs).
BBW: Let's talk about your own investments, outside of what you're doing at Kleiner. My list says Smartpipes, Kibu, which went down already, MyCFO, and Shutterfly. Am I missing any?
TJ: There's another one, called IP Unity. They're building edge devices for voice over IP. Stuff like hardware concentrators and software applications to enable commercial-grade phone.
BBW: Who's the target customer for something like that?
TJ: Anyone doing infrastructure for communications businesses. So the Qwests, the Level 3s, the WorldComs. The local CLECs (Competitive Local Exchange Carriers) and DLECs (Data Local Exchange Carriers).
BBW: What I'm hearing is that lots of people have pieces of the VoIP puzzle, but nobody really has the total package, that takes an IP call from one end to the other, and all the parts in between.
TJ: Yeah. Well, at Kleiner, we're pretty heavily invested in startups that provide different pieces of that puzzle. IP Unity is an example of one of them that we invested in to build this one piece of it.
BBW: What's hot on your list for KPCB investments?
TJ: We've got companies that are in free-space lasers, for example. Or, I should say, a free-space laser startup. Another is a high-speed CMOS (complementary metal oxide semiconductor), that drives fiber out of the CMOS. And one's a company that off-loads TCP/IP and queuing for high-concentration, backend database applications.
BBW: All fairly infrastructure-related, sounds like?
TJ: You know me. I'm almost always involved in core infrastructure.
BBW: What about broadband trendlines in general? You're once again in a pretty good catbird seat for this kind of conversation.
TJ: I'm still a huge believer. The only reason I'm working at all is that I still have quite an interest in broadband. In four years at @Home, I learned that there are still tremendous pieces of this puzzle that haven't yet been addressed, as far as getting true broadband to consumers and businesses.
BBW: Such as?
TJ: We had one approach with @Home, which was, screw it: We'll build our own. We were way ahead of our time. We were smart to do it that way--I'm not trying to pat my own back here, but, the fact is, everyone's trying or has tried to duplicate the model. Thing is, (to do that) you need proxies, data centers, network operating centers. By building it, we were able to control a lot of the equation.
But on public networks, you don't have the same capabilities, so you have do things in different ways.
BBW: You mean to do low latency, high bandwidth services on the public Internet, versus the benefits of a managed, private backbone?
TJ: Right. So most of the companies we're involved with, and all since I've joined (KPCB), are companies focused on how to speed the end-to-end equation, to broadband homes. Stuff that gets you to the point where can have real-time video streams, and video time-shifting, and music time shifting, and truly high quality, commercial voice. You need super low latency, and super high bandwidth, all the way from one end to the other.
BBW: And unless that happens...?
TJ: And unless that happens, without the next generation of technology, broadband won't deliver on the promise. It really is starting to get there. I mean, 4 million cable modem customers at the end of 2000--that's starting to be significant. So over the next couple of years, as the number starts to get to be 10 million, 20 million, the rest of the infrastructure will collapse around it.
BBW: No matter how much dark fiber there is out there?
TJ: No matter. Today, less than 10 percent of the buildings that house over 200 (people) have fiber going to them. Worldwide, it's less than 1%. So what are you going to do about it? Are you going to rip up every street and lay fiber? (laughs) That's why the area of free-space lasers is such a big deal. That's one part of the equation. With VoIP, it's making it real--not a hobby. Being able to drive fiber out of the CMOS means you can stay in fiber, and not have to go into the electronics domain, cost effectively. So, each project is aimed at end-to-end high-speed broadband, with low latencies.
|
