Skip to Content

Bankrolling the Future

Ethernet inventor Bob Metcalfe and Nobelist Walter Gilbert, now both venture capitalists, meet to pick the most significant emerging technologies in IT and biotech.

Earlier this year an intriguing coincidence took place in the world of innovation: at almost the same time, Bob Metcalfe and Walter Gilbert decided to become venture capitalists in the Boston area. Both Metcalfe and Gilbert are distinguished innovators who have played many different roles in the process of bringing research to market. Metcalfe invented the Ethernet and founded 3Com to commercialize it. After leaving 3Com, he went on to a career in publishing and punditry. Gilbert has had a brilliant research career in molecular biology, sharing the 1980 Nobel Prize in chemistry for discovering a fundamental method for sequencing DNA. He cofounded one of the very first biotech companies, Biogen, in 1978 and has also had a hand in starting several other biology-based companies, including Myriad Genetics.

Now both men are approaching the world of entrepreneurship and innovation from a new angle. Metcalfe has stopped writing his column in InfoWorld and joined Polaris Venture Partners of Waltham, MA. Gilbert has taken a leave of absence from Harvard University to join BioVentures Investors, in Cambridge, MA. In spite of the parallels between their careers, Metcalfe and Gilbert come from different worlds-information technology and biotechnology-and had never met. Technology Review thought it would be interesting to bring them together to talk about the current climate for investment, the most important emerging technologies and why they became venture capitalists now. Editor in chief John Benditt mediated a conversation over dinner in downtown Boston.

TR: How is venture capital activity different from when you started out forming new technology-based companies?
Gilbert: Back in the 1970s, venture capital groups were trying to find a few engineers or a few scientists to put them together with business people and form a company. There’s very little of that now. To do that takes great effort, constant nurturing, and there’s only a small amount of money involved at that stage. So if one runs a very large fund, one can’t afford to put capital into small “seed” efforts like that because that would spread the money into too many companies. Nowadays, when a venture capital group supports the growth of a small company it is generally in the second, third or fourth round, with large investments, as opposed to the very beginning.

TR: How has the broader picture changed since you founded Biogen?
Gilbert: We have shifted from a period in which people doing biology had an ivory-tower view of the world, in which there were no practical applications for what they did, to a view in which everything appears against the background of immediate practical application. Back then, it was like breaking your oath, breaking the code of purity, to think about founding a company. There are thousands of biotechnology companies now; there were none then.

TR: How was it possible for you to go beyond the idea of an “oath of purity” and get involved in founding a company?
Gilbert: That results from having a strong view of the public good and believing that molecular biology could actually make a useful pharmaceutical, so that it was worth spending my time to do that and to show that it could be done. At that time the rest of our community didn’t think it was possible. I learned a lot about the world doing this that I didn’t know before, and in some ways, serving Mammon is cleaner than not serving Mammon, because greed is a very simple motivator, and other motivators in science are much more complex.
Metcalfe: But there is an ideology question floating around here. There are people who believe that making money is evil. They’re wrong. But they believe it.

TR: Right. So how were the two of you able to get beyond those blinkers?
Metcalfe: Well, if you want to see your ideas out in the world, what is the most powerful way of projecting them? Not through a nonprofit organization begging for donations. The best way to convey ideas, I think, is with the people who had the ideas. I don’t think it works well for the academics to sit here and generate the new ideas and then write the papers and then have other people put them into practice. I learned that at the Xerox Corporation in the 1970s.

TR: What was your experience there?
Metcalfe: Well, we had this model that there would be a research center, and then there would be this advanced academic development laboratory, and then there would be engineering departments, and there would be these products. So the guys in the science groups, where I was, kept throwing these ideas over the transom, and the guys on the other side either kept dropping them, or they were proud of their own work, so they would ignore everything we did and start over. But they would come up with what we considered to be dreck. So a few of us started migrating downstream, taking the ideas with us instead of trying to get them adopted by throwing them over the transom.

TR: The two of you have been both basic researchers and entrepreneurs. Now, in your role as venture capitalists, do you look for both those capabilities in one person?

Gilbert: We seldom look for that. It’s too hard to find. A more typical pattern in a startup company is that one has a technical group, including the people who came up with the idea, and also someone else who’s trying to push the business, worrying about the money-raising, creating the capital.

TR: What is the investment climate like now in venture capital?
Metcalfe: A tremendous investment stream has been created to about five or six hundred venture capital firms, about three or four hundred of which are completely screwed up, and most of them are going to die. And the limited partners, the pension funds and so on, who provide the funding for the venture capital firms are saying, “Okay, we’re going to invest in the two hundred firms that didn’t screw up.” But that can actually create a problem in that the limited partners want to give your firm more money than you can reasonably deal with right now.

TR: Some would say it’s a funny time to become a venture capitalist. We’ve just been through this great big “new economy” bubble, and everybody’s feeling burned. So why did you two distinguished innovators decide to try it now?
Metcalfe: It has to do more with my personal life cycle than with the state of the markets. I had finished my previous career as a journalist, and venture capital was the next thing. I like entrepreneurs and I like high tech. This is another way-as journalism was before-to stay with those topics and those people. Just another way. New and challenging things to learn but still pretty much involved with innovation.
Gilbert: Well, actually, this is a great time to be a venture capitalist. Valuations are low, so it’s an excellent time to buy a company. Too many people think that if valuations are low, that means they’re going to stay low in the future, as opposed to “valuations are low today, but the cycle will take us back up again.” In my case, I’ve been very interested in small companies. I’ve founded a number of small companies in recent years, and so I stepped back and said, “Well, maybe I shouldn’t just ask for money, I should participate in the venture capital structures that give money to these companies.”
metcalfe: I heard this from a man named Paul Deninger, who runs Broadview, which is a major mergers-and-acquisitions company, and his advice was, “Forget what happened over the last two years, because there’s nothing to be learned from it.”

TR: Nothing?
Metcalfe: Nothing. It was just a crazy aberration, and if you try to learn something from it you will be misled and distracted. Delete it from your memory, because it was just such an extreme speculative bubble. The trouble is, there’s now an army of zombies whose thoughts about how the world works were formed during this period. They’re doomed, because for the rest of their lives, they will try to re-create that situation, which won’t recurat least for a long time. You see them occasionally when they come in, expecting a pre-money valuation of $175 million because there’s nine of them and they’re really charming.

TR: So the recent experiences don’t inform what you’re doing as venture capitalists? They don’t change what you look for?
Gilbert: If you’re serious as a venture capitalist, you try to look at the serious aspects of companies. Do they have real intellectual-property elements which they have protected that could be of commercial value? And if there is commercial value, what is the nature of that value? The dot-com cycle was a strange one, because people were interpreting the number of clicks on your Web site as income.
Metcalfe: The basic idea of the bubble was, “We’re going to give away what it is we do best and make money some other way, such as monetizing the eyeballs that we’re able to attract.” You see this with Linux, where they say, “We’re going to give away our software, but we’re going to make money some other way.” The trouble was, those other ways never developed, in particular the anticipated advertising never materialized.

TR: Now I assume you [Metcalfe] are working mostly in information technology and you [Gilbert] mostly in biotechnology. What are the areas you think are most exciting these days?

Metcalfe: I don’t think those are two different categories. I think it’s a continuum. This is an argument I have with my partners. The way my partners characterize our investments is that we’re two-thirds info, one-third bio. But I’m a curious person. I love to learn. So I sit in on these biotech presentations. And the bio guys are talking about computers and bits and databases, and, of course, many of the computer guys are talking about applying biology to computing. So I think it’s healthier to consider it a continuum, rather than this is info, this is bio.

Gilbert: It is a bit more confusing than that in biotech. There are several different types of companies. Obviously, one type of company is developing novel products, novel drugs. There’s another whole field of companies developing novel technologies, which can help other companies develop products. Companies like Celera, for instance, which is sequencing the human genome. And then there’s the whole field of bioinformatics, the issue of whether you can apply computers to biology. The amount of information you’re trying to work with can only be dealt with by information technology, and the questions you ask can only be phrased in terms of computer programs and computer analysis.

TR: We’ve had a wave of genomics companies, and the next phase seems to be proteomics, companies trying to identify all the proteins in the human cell and the ways the proteins interact.
Metcalfe: I’m involved in a company that’s the next wave after that. There’s genomics, and there’s proteomics, and there’s the next one after that. It’s called glycomics, and I’m going on the board of a company that does that (see “Glycomics,” p. 33).

TR: Tell us about it.
Metcalfe: Well, the DNA creates the proteins within the cells, and then sugar molecules are added to the proteins. The sugars affect how the proteins behave. The alphabet of sugars is more complicated than the alphabet of proteins, which is more complicated than the alphabet of DNA. Please check me if I’m wrong.
Gilbert: Well, there’s a slight exaggeration here. I don’t want to disparage your company, but up till now the glycosylation patterns, the patterns of sugars attached to the proteins, has been a technology desperately seeking a useful target.
Metcalfe: This came about because Bob Langer of MIT, who has done probably 10 or 15 deals with Polaris, comes in with two PhD guys who have done this glycomics work. So I asked whether it’s possible that cancer cells attach their sugars in a specific way. And they said, “Yes, in fact, we think we might be able to detect cancer way early by looking for these specific patterns of glycosylation,” or attachment of the sugars to the proteins.

TR: Dr. Gilbert, is this also an area you’ve been looking into?
Gilbert: No, although given an opportunity I might invest in that company.

TR: Is this postgenomics?
Gilbert: Well, actually it’s both pre- and postgenomics. Glycosylation has been around for a long time. And yet no one has yet shown that there is a real alphabet of glycosylation that could be deciphered as the genetic code of the DNA has been deciphered that would prove to have practical value in terms of how the proteins behave.

TR: What would you pick as upcoming significant technologies in your fields?
Metcalfe: I am telling people the next big thing, the next big thing on the Internet, is video.

TR: Software to convey video over the Internet?
Metcalfe: Well, the bandwidth. The content, the switching, the bandwidth, everything related to the delivery of video.
Gilbert: Perhaps. But there is still a major question of whether this might develop in a peer-to-peer fashion or as a broadcast medium, with a central source sending content out to large groups of viewers.
Metcalfe: Yes, well, we had the same experience with earlier stages of the Internet. What you saw is that communication led and content followed. E-mail beat out weather and stock quotes. They came later. So the question you’re asking is, will the initial killer apps of video be communication or content oriented? Will they be movies, or will they be videoconferencing or video telephony? And if there’s a recapitulation, then you’re absolutely right, the videophone will drive video rather than video-on-demand movies.

TR: Sure, but lots of people have said video on the Net was the next big thing. So why does your gut say that it will actually materialize now?
Metcalfe: Well, in microprocessors, we went from one-bit to two-bit to four-bit to eight-bit to 16-bit to 32 to 64. Bill Gates got to be the wealthiest man on earth because he knew this progression would continue. And he made his money on the transition from eight to 16. A lot of people said, “Well, we’re not really fully utilizing the eight-bit microprocessor, and there’s no software for the 16-bit microprocessor, so it’s ridiculous that anyone would bet on that.” And Gates knew 16 was next and 32 was after that. Now, in the bandwidth world, the same thing is happening: one kilobit, 10 kilobits, 100 kilobits, 1,000 kilobits. It’s sort of an obvious progression related to the unfoldings of bandwidth. The backbones are now 10 gigabits, but it’s going to 40 gigabits.

Gilbert: And what do you see as the underlying infrastructure?

Metcalfe: I’m on the board of a public company called Avistar [Communications] that does desktop-enterprise videoconferencing. This dates back to before my involvement with Polaris.

Gilbert: Well, we’re still waiting for good videoconferencing. We’ve been waiting for years, and there may not be good videoconferencing for another 20 years.

Metcalfe: Well, this company is the next attempt. I’m watching them introduce new versions of their software with new features, and then we all watch with anticipation at the usage levels, measured in minutes per day and how many users there were and how long it takes to sell the new feature. The sales cycle is long because there are so many people walking around who have had their hearts broken in previous versions of videoconferencing.

TR: What else will be needed to make the video revolution go?
Metcalfe: Well, there’s a company I can’t name but whose business plan is before Polaris. They make a chipset for editing HDTV. And there isn’t a lot of HDTV around now.
Gilbert: Not a lot of broadcasting yet.
Metcalfe: Yes, but we think that HDTV is the coming thing.
Gilbert: And your basis for thinking that?
Metcalfe: The inexorability of progress and the fact that HDTV is gorgeous. More people will want it eventually. And it’s not enough to have the TVs, and it’s not enough just to have the broadcast facilities. You’re going to need the editing facilities. Therefore you need these chips. The consensus among the partners is that we should probably do this investment, because these chips are going to be needed. My own attitude was triggered by the fact that my 12-year-old son gave me as a Father’s Day gift a movie that he made using [Apple Computer’s] iMovie on the Macintosh, with a digital camera, titles, music. So this is coming.

TR: Are the chips for professionals or consumers?
Metcalfe: The initial plan is based on professionals. As the prices come down, this will become a widespread phenomenon; just as people write letters, they can make movies for grandma and send them over the Net.

TR: If your grand vision is correct, does it matter to you when it comes true?
Gilbert: It’s very easy to pick things that will prove profitable. The question is, when? The entire issue is, can you invest in a way that produces a return very very soon?
metcalfe: Timing is everything.
gilbert: Timing is everything. Timing is every element of the next strong investment.

TR: Dr. Gilbert, Bob Metcalfe has gone out on a limb and said video is the next big thing in IT. What do you think is the equivalent from within biology?
Gilbert: I don’t think I know anything that’s going to be as important as Bob thinks video is going to be. In biology now we have some very large fundamental problems that need to be solved, problems of basic scientific knowledge. We now have a very good view of the human genome. But we don’t have good ways of understanding how many genes there are in the human genome or how many relevant proteins there are. That is, in fact, a bioinformatics problem, and it hasn’t been solved.

TR: So one issue is understanding how many genes there are?
Gilbert: We can’t do that yet in a totally convincing way. I would think we would be able to within a few years. And I think we will in fact also work out the two next major problems in biology. One is the problem of how proteins fold up into their specific three-dimensional shapes-how to predict that from their DNA sequence. And the other problem is deciphering the program of how the egg develops into the complete organism.

TR: And are those problems being solved by the companies you invest it?

Gilbert: The scale of these problems is beyond the reach of the average small company. They’re very large problems. And if they’re solved, then it’s difficult to sell the result. Is it useful for a company to try to solve the protein-folding problem? In some ways it’s questionable. It’s very hard as a company to sell that as a useful approach. There are companies trying to do that. Would I invest in them? Probably not.

Metcalfe: Well, I’m involved with a company that uses peer-to-peer software for doing protein folding. And it seems to me that there is a way to make this kind of work profitable. Couldn’t you just run through the human genome, which has now been catalogued, and fold every protein, and create another database of the foldings of every protein implied by the genome?

Gilbert: You could in principle, but you can’t in practice, because you can’t do the calculation.

Metcalfe: Well, let’s assume that in the fullness of time, they will solve the problem, so now we have a complete database that is not the human genome, but it’s all the expressed proteins of all the genes of the human genome. What could you do with that?

Gilbert: Well, you could sell it to somebody. But can you sell it for enough to have made the whole thing worth doing? It’s like with Celera. Are the companies they sell their DNA sequences to going to offer enough to make the half-billion dollars in Celera’s costs to do the genome worthwhile? Are they actually going to get enough income to match that? I don’t think so.

TR: So those problems will be solved in the universities or by government funding?
Gilbert: Well, the big problems, the fundamental problems, get solved in the university. And we need the government-funded projects for things that are far out. Celera would never have been able to sequence the human genome without a massive initial public investment.

TR: So does that mean we have to wait until these huge problems are solved to get a return in biotech?
Gilbert: No, not at all. The individual drugs will be profitable, not the whole database of information. In many areas we can already find specific molecules that will have important effects. I mean, the exciting subfields are things like neurobiology-can one enhance memory? I’m involved in a company trying to enhance memory, so I think that’s a perfect example. With Eric Kandel of Columbia University I helped to found a company called Memory Pharmaceuticals.

TR: What is the company doing?
Gilbert: They develop small molecules to use as drugs that affect memory. We already know enough about what goes on in the hippocampus [a brain structure that plays a central role in memory] to enhance the way that the cells talk to each other. And one can study the process of memory further by turning genes on and off in the hippocampus of mice. You can make dumb mice smart and vice versa by switching the right genes off and on.

TR: Well, thank you both very much for speaking with the readers of TR. One of the things that’s most interesting to me in this conversation is that although you come from different backgrounds-in infotech and biotech-there is so much overlap in technologies these days that you may wind up in your new role as venture capitalists competing for the same companies.
Gilbert: Well, my general reaction to the discussion we had is that we may be potential coinvestors.
metcalfe: It’s a win-win situation.

Keep Reading

Most Popular

Large language models can do jaw-dropping things. But nobody knows exactly why.

And that's a problem. Figuring it out is one of the biggest scientific puzzles of our time and a crucial step towards controlling more powerful future models.

OpenAI teases an amazing new generative video model called Sora

The firm is sharing Sora with a small group of safety testers but the rest of us will have to wait to learn more.

Google’s Gemini is now in everything. Here’s how you can try it out.

Gmail, Docs, and more will now come with Gemini baked in. But Europeans will have to wait before they can download the app.

This baby with a head camera helped teach an AI how kids learn language

A neural network trained on the experiences of a single young child managed to learn one of the core components of language: how to match words to the objects they represent.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at customer-service@technologyreview.com with a list of newsletters you’d like to receive.