Interview: Drew Lanza, Morgenthaler (For Now)

Drew Lanza puts the future of the chip sector, the investment potential of Internet application software, and rutabaga soup on his lunch menu

Craig Matsumoto, Editor-in-Chief, Light Reading

June 7, 2010

11 Min Read
Interview: Drew Lanza, Morgenthaler (For Now)

Drew Lanza, a partner at venture firm Morgenthaler and a contributor to Light Reading since its early days, will soon be out of a job. His firm, one of the strongest supporters of telecom startups, is turning its eye (and wallet) toward applications instead. (See VC Focus Turns Away From Chips, Systems.)

Morgenthaler is far from alone, and the trend -- fewer venture dollars for chips and systems -- could have large-scale ramifications, particularly in Silicon Valley. Lanza and Light Reading hammered out some ideas over lunch at Madera, the year-old restaurant located right in venture capital's Sand Hill Road enclave.

Lanza, over a bowl of soup and a Caesar salad, recalled his favorite historical parallels for the tech boom, some of which have only gotten stronger. We also discussed what happens in a world with fewer chip startups -- who does the inventing?

In the end, we didn't come up with many answers, other than to determine that rutabaga soup is better than it sounds. We're posting this to show the questions that arose, and to solicit yours. If this change in Silicon Valley is not cyclical -- if the age of chip startups is passing -- what does that mean for technology, and even society?

— Craig Matsumoto, West Coast Editor, Light Reading

Light Reading: I can accept that right now is not a good time to start a chip company -- there's a lull in interest. But then, where do the new innovations come from? Someone has to start a chip company somewhere.

Lanza: There's a flip side to that. My partners say we're going to do Internet services, and my comment in the long-running debate is: Well, that's all right, but if everybody says that, then that becomes the tragedy of the commons. All the sheep are grazing on one field and pretty soon, there's nothing left to graze on.

If I were a senior government person in the United States right now, I would be very, very worried. You've seen those articles: Entrepreneurs -- specifically, venture-backed entrepreneurs -- have created something like 70 percent of the net new jobs outside of the government sector in the last 20 years.

Light Reading: And besides, how much money can a VC firm make off of Internet services? Not all services are big and expensive.

Lanza: It doesn't take a lot of money to go into it, which is attractive.

As an industry, we raise $20 billion a year. In good years, we raise $30 billion; in bad years, we raise $10 billion. Say each one of those startups takes $20 million over its lifetime. That just doesn't work. That says we need to start 1,000 of those per year? And make them profitable? I don't think there's that much energy out there for Internet services.

So, that's the view from 100,000 feet. You must be seeing a lot of this as you talk to people. It's pretty tough out there.

Light Reading: It is. You can see it in the news. All the funding, all the excitement, is about the services. Or iPhone applications.

Lanza: I always think of these analogies, in terms of the Industrial Revolution. It's very clear that the iPhone, the smartphone, Android, are to the information revolution what the Model T was to the Industrial Revolution: that thing that everybody uses that changes their lives. And interestingly, per that analogy, we've just hit a big depression or recession, partly out of the boom created by that. Hmm, OK. So, part of the analogy's holding, right? We're witnessing large government upheaval -- well, that holds as part of the analogy. There were other industries formed after that. The airplane industry was notably formed in the '30s, '40s, and '50s. People made a lot of money doing that. I don't know what the equivalent of that is in our era.

Light Reading: That makes sense. It does feel like the iPhone is something that leads to other things being invented.

Lanza: I just can't believe we're really done with hardware. We've been building cars since the Model T, but as we've seen from the bailouts, that's become a pretty crafty industry over time.

So, maybe, yes, we're always developing and designing new hardware, but it's not a very exciting place to be, and maybe it gets consolidated into large Fortune 100 companies. Cisco Systems Inc. (Nasdaq: CSCO) will always do 10 or 20 new chips a year. Huawei Technologies Co. Ltd. has an arm that does chips. And certainly the whole thing with FPGAs -- we'll always have some method for instantiating electrical property into hardware, whether you do your own ASIC or do your own FPGA or program a microprocessor or microcontroller.

Light Reading: With what you can do in FPGAs now, with the horsepower you get in microprocessors -- maybe I can talk myself into believing that we don't need a chip industry anymore.

It's hard to fathom. It's hard to believe there isn't going to be something new that comes up that requires some orthogonal path.

Lanza: Yeah, but those are always the interesting ones, the ones you don't see coming.

Light Reading: I guess you can't raise millions of dollars and sit around waiting for a few of those.

Lanza: No. Part of what I want to do this summer -- I've kept good relations with most of my professor buddies at Stanford over all the years. I want to spend time buying them beer -- and if you know your Stanford professors, they love beer -- pick their brains, and see what's interesting to them.

The inevitability of Moore's Law -- I think it's called Arthur Rock's Corollary -- is that sooner or later, there's only two fabs on the planet. We're probably within a decade of that.

So, how does that work? Does everything become software? Andy Rappaport, who's a really smart guy, wrote an article many years ago, "The Chipless Chip Company," sort of as a reaction to fabless chip companies. Pretty soon, they're not going to do chips either. I think he wrote that in the mid-'90s, which is very prescient.

Light Reading: What happens to systems design, then? That one I can understand more: You may need new systems, but it takes so much money to build them, and everyone lost so much money on so many of them -- that one, I wouldn't get into anyway.

Lanza: Do they just get built by Cisco? Do Cisco and Huawei just get bigger and bigger and bigger?

Light Reading: How depressing is that?

Lanza: It looks like the automobile industry, in that case. There's a hundred car companies, and after a while...

Light Reading: ...it's down to three. [Ed. note: Two, actually.]

Lanza: Some fools down the road here are trying to start a car company. How's that going? If the government wasn't giving Tesla lots of money, they wouldn't be going. So, maybe to some extent that analogy holds -- that there are only a couple of big suppliers, and that's it.

Does that mean we're at the end of a massive wave of consolidation where industry and government gets bigger, and the little guy just gets squeezed out? Because we're not starting startups in hardware. It's getting very tough to start restaurants and banks. The government's growing by leaps and bounds. If we have to keep building new systems but startups are doing it, that means industry -- Fortune 500 -- is going to grow.

We saw that in the 1950s, with the creation of the first big conglomerates like ITT. Maybe we're back at that stage.

Light Reading: That would be really interesting.

Lanza: Everything's just big for a while. Until people just get fed up with it.

Light Reading: It's happening in a lot of industries -- the Wal-Mart effect, the Barnes & Noble effect.

Lanza: It's happening everywhere. Even Huawei in China.

Silicon Valley was built on entrepreneurship. Obviously, we'll still have stuff going on in Internet services. But it seems to me like a whole big leg of the stool is about to crumble.

My kids are all studying bio-engineering. Well, two out of three. The other one's studying wine tasting -- she's my favorite. Smart kids, because in 20 or 30 years, I think that's inevitably the next big thing. But there has to be a bunch of stuff between here and there.

Maybe it is all Internet service. Maybe everything is about software that people can work with -- not software in and of itself, but software deployed as a service.

Light Reading: Or, maybe there's a pause until we can exhaust everything we can do with the Internet. Maybe right now, there are just so many possibilities open...

Lanza: Wild West, right?

Light Reading: ...that software is all that matters? You know what, that's not true. It's just that no one wants to put money into hardware.

Light Reading: It could be that there's a long period where we don't need any radically new hardware, because we're still working with the stuff we've got -- hardware has exceeded the capabilities we've got. There's a gap.

Lanza: Well, and that follows the industrial model, sort of. When you drive a car today, it's really not that much different from a Model T Ford. It's an internal combustion engine, driven by gasoline, runs on rubber tires, right? It has a steering wheel, modeled after a ship's wheel. And a Model T Ford can go 30 miles an hour, and if you go much faster than 80, the cops'll stop you. So, with 100 and some-odd years of progress, we go twice as fast as we did before, and 100 companies have been whittled down two. Maybe systems are ready to be checking out.

Although, that doesn't quite jibe with this whole notion that bandwidth demand is an exponential. Maybe the Internet does slow down. Maybe once you get 100 channels of video, you declare victory and say that's it.

Light Reading: Yeah. There is an upper limit. There's got to be an upper limit.

Lanza: The human brain can only absorb so much, and machine-to-machine communications -- I think someone said they're half the traffic on the Internet, but I think they're just moving bits around.

Light Reading: It's a huge volume of little tiny siphons. A machine's not going to watch Die Hard over and over again.

Lanza: No, but it might move 100 copies of it around, without somebody in the middle.

I wrote that article that within our lifetime, that the majority of bits would be carried at 10 Gbit/s, that we would hold that bit rate for two or three decades. Maybe I'm right. Sure, there will be 100-Gbit/s serial links built. But for the bulk of bits, does the Internet become like the Model T Ford? Roads are the same as they have been forever. Maybe we have new technology in how we lay out an asphalt slurry, or we put the road turtles on so we don't cross the lane, but essentially you and I are driving vehicles on roads with rules that go back 100 years.

Light Reading: I kind of agree with that. The Internet gets to the point where it's just like the phone network. It already does everything you need it to do -- except you get to work on it yourself. You get to tweak it, add features, do HTML5 -- but it becomes a basic utility where the purpose it's set up for, it can do.

Lanza: The maximum bandwidth you need per human being is about 50 Mbit/s. Every time I put that down, I get this massive debate on Light Reading: "No, it's not! It's like a gigabit!" No, that's uncompressed. The human brain deals with compression. If I can fool the human eye, then by definition, I don't need more bits, right?

I think you can give somebody both eyes, full field, 50 Mbit/s, and they couldn't tell the difference between reality and that.

Light Reading: And compression will get better. People aren't just going to let these bit streams keep growing.

Lanza: So, that's just not that many bits.

So, what happens to all of us, then? We've been doing this for decades. I've gotten calls from people for the last decade saying, "Hey, Drew, help me find a job." I've probably placed 100 of my buddies for the last decade. But I tell them all the same thing: "Pat yourself on the back, go have a beer: You built the Internet," you know? "Stop complaining."

— Craig Matsumoto, West Coast Editor, Light Reading

Back to the Introduction

About the Author(s)

Craig Matsumoto

Editor-in-Chief, Light Reading

Yes, THAT Craig Matsumoto – who used to be at Light Reading from 2002 until 2013 and then went away and did other stuff and now HE'S BACK! As Editor-in-Chief. Go Craig!!

Subscribe and receive the latest news from the industry.
Join 62,000+ members. Yes it's completely free.

You May Also Like