Home » Uncategorized » Photon Man

Photon Man

You’ve heard of Moore’s Law, where semiconductor price/performance doubles every 18 months; now get ready for Kalkhoven’s Law, where bandwidth availability doubles every 12 months when optoelectronic technologies are deployed. Kalkhoven’s company, Uniphase, which is positioning itself as the leading supplier of optoelectronic components, shot up to eighth place on this year’s CE Growth100 ranking.

When Bill Gates said that “we’ll have infinite bandwidth in a decade’s time,” some said it was another prediction wide of the mark. They were half right. Unlimited bandwidth may be available much sooner. The ubiquitousness of computing, video, interactive services, and multimedia-not to mention the Web-has placed enormous demands upon the network communications infrastructure. Legacy networks of the telcos are under pressure to transmit massive volumes of information. Data traffic is doubling every 100 days, according to the U.S. Department of Commerce, and Internet use is rising tenfold each year despite the limitations of access bandwidth (see graph). As former FCC Chairman Reed Hundt said, “What we need is a data network that can easily carry voice, instead of what we have today, which is a voice network struggling to carry data.” The appetite for bandwidth continues to fuel debate over the price and performance of new packet and cell-based architectures.

According to industry observers such as author George Gilder, who publishes the Gilder Technology Report, the emergence of new optical technologies will permit bandwidth to be come so cheap that it may reduce the need for both switching and power, “thus reversing the calculus of abundances and scarcities” based on silicon. Owing to its emerging leadership as a supplier of advanced optoelectronic components, the San Jose, CA-based Uniphase is positioned, in Gilder’s view, to become “the Intel of the Telecosm.” Not bad for a company founded in 1979, which in the early 1990s generated a mere $30 million in annual revenue as a maker of helium neon lasers for portable bar code readers. In 1992, the board of directors hired Kevin Kalkhoven, 54, an exuberant Australian and 25-year veteran in the computer industry who had been president of Demax Software and CEO of Aida Corp., to invent a new future for the company.

For the Adelaide-born former COBOL programmer, the epiphany came in January 1994, when MCI installed a 2,000-mile fiber optic line between Chicago and Sacramento with no repeaters. (The use of erbium doped fiber amplifiers permitted the breakthrough. Fiber signals normally require repeaters and are subject to degradation.) Kalkhoven saw that a confluence of new technologies, among them wave division multiplexing (WDM), promises to create an all-optic standard that overcomes such transmission limitations. (WDM increases the capacity of optical fiber by simultaneously transmitting signals at different wavelengths through the same strand of fiber.)

Kalkhoven set about acquiring companies to assemble an integrated optoelectronic capability. Critical pieces came together in 1997 and 1998 when he acquired IBM‘s semiconductor laser plant in Zurich and Philips’s Optoelectronics facility in Eindhoven. But to capture a commanding position in this ascendant technology, Uniphase needed to master fiber gratings, which are used for filtering and selecting wavelengths. Using his frequent flyer miles, Kalkhoven hopped on a plane to Sydney to meet Simon Poole, who had left British Telecom to pioneer holographic gratings and other optic components for a start-up company he founded called Indx. Over beers at the Sydney Opera House, Poole agreed to join Indx and Uniphase.

According to Hambrecht & Quist’s Jeffrey Lipton, optical networks greatly expand the cost/performance of communications. He sees Uniphase as a leader in a still nascent industry competing with companies “that have significant technical expertise and financial mass including vertically integrated vendors such as Alcaltel, Lucent, and Nortel, and smaller focused companies such as Ortel and SDL.” BT Alex. Brown principal Chris Crespi, a former optics engineer at H-P, reckons the Uniphase and Ontario, Canada-based JDSFitel merger will give the new combined company the switching capability necessary to stay ahead of the curve. “Uniphase has a strong value position,” he notes, adding that they will need to broaden the product line and continue to push the envelope with advanced technology.

Chris Crespi thinks it will be a few years before people can expect optical fiber in their homes. “For one thing, you have to find a way to cut and splice it and stick a jack in the wall,” he notes, pointing out that the technology is in its early stages of evolution. “Don’t discount Japan, Inc.,” Crespi adds. “They missed the boat on this, but look how they came back in CD lasers.”

Kalkhoven has everything riding on the deal with JDS-Fitel, which is valued at $6.1 billion if the merger of equals comes off. The deal will make the combined company the largest supplier of active fiberoptic components and modules in the world. JDS Uniphase, as the new firm will be called, will operate in eight countries with pro forma sales of about $420 million this year. Kalkoven will become co-chairman and CEO and Jozef Straus from JDS-Fitel will become co-chairman, president, and COO. Both companies will name half the new board. “It’s a big bet, says Kalkhoven, “because each of us has been successful separately. We could put them together and screw it up, but we’re working damn hard to make sure that doesn’t happen.”

You predict the transition from a computer-centric world to a bandwidth-centric world. How will that happen?

n two or three different ways. First, of course, through the advent of the Internet where the channels of communication have become almost boundless. The problem is that it’s slow. It’s not the World Wide Web. It’s the worldwide wait. Second, high-speed bandwidth will enable new industries in the same way the microprocessor did. Until the microprocessor there was no AOL, no Intel, no Microsoft. Fiber optics and bandwidth will create new industries that we can’t even think about today purely because instant access to information and people will change the way we think about both society and the way we work.

Just recently, people used hand-held video cameras and the Net to send video from Kosovo and Croatia. That simple set of communications changed the information we were getting and the way we thought about those places. So, that’s the big opportunity of the next decade.

What will suddenly make bandwidth as available and cheap as running water?

Two technological advances-optical amplifiers and semiconducor lasers-will enable unlimited bandwidth. We’ve had fiber optic technology before, but you had to electronically regenerate the signal every 40 kilometers or so. In 1988, at The University of Southampton in the U.K. they invented something called an optical amplifier, which lets you put in multiple wavelengths of light, each of which carries as much as that fiber carried before.

The other technology is semi-conductor lasers able to handle this communication without breaking down. These two key elements were then put together into what we call wavelength division multiplexing or multiple wavelengths of light-the ability to carry multiple wavelengths of light down the fiber.

How does that compare to the capacity of dumb copper wire?

Take all the data and telecommunications traffic in the U.S. over a 24-hour period, reduce it to one second and that’s what we can carry. We’ll be able to go from 2.5 gigabytes per second, which would be one piece of fiber, one wavelength, to 80 wavelengths in each at 10 gigabytes per second. And we’re at least doubling the performance of our stuff every 12 months. Imagine going from a modem speed of 56 kbps up to two or three megabytes.

That sounds terrifically expensive.

It’s actually phenomenally cheap, because you’re using the same infrastructure, the same fiber optic networks that currently exist. But instead of those networks having to amortize their cost over 2.5 gigabytes per second for an increase in capital cost of 5 percent to 10 percent, you can increase the amount of data traffic that they can take and therefore sell by 50 or 80 times. So, for a small increase in capital cost you’ve got an eightyfold increase in potential revenue. And, because of the economics, that translates into a decreasing price for the consumer.

How does this relate to Internet telephony?

A typical telephone conversation between New York and San Francisco involves opening a circuit that ties up lots of switches, fiber, and open capacity. Yet the amount of traffic actually going down that line is extremely limited. The quality is excellent, but it’s a very expensive way of communicating. With IP or Internet protocol, you move everything into packets of information so your voice and 100 other voices all get mixed up on the same packet and sent down the line and you assemble and disassemble the packets of information at each end.

It’s clearly a cost-effective way of communicating. The problem is that the routing technology has not been fast enough for the quality we’re used to. It’s fine for data, but the nuances of inflections and tones have been difficult to reproduce. But we are very close to equipment that can assemble and disassemble at phenomenally high rates. When that happens there is zero doubt in my mind that we will see voice over IP.

Then it will be a five-cent call to anywhere in the world, which will change the way we think about this global society. One, people will use this stuff more. Two, we’re going to see growth in industries that we can’t even think of today. We are on the edge of a whole new set of industries that may not even exist today, but will be major, multi-billion dollar players in the next 10 years.

What do you think we might actually see?

The ability to remove the element of consumer’s time is really what high-speed communications means. Already, we’re seeing the potential demise of travel agents as people book air travel directly on the Net. Clearly we see the opportunity to revolutionize not only the way in which we shop and work, but also things like tele-medicine.

The way medicine operates today, it consumes consumer time rather than the limited resources of medical practitioners. Seeing a doctor is a full-day exercise. With high-speed communications, there is no reason why you couldn’t have a lot of remote computing-based medicine. You should be able to go to your PC and effectively speak to your GP in a real-time environment. You could have banks of remote clinical operators talking to people in their homes and doing medical diagnosis.

You can see similar effects in education. It’s a bizarre society that demands that you gather people and try to educate them all at the same speed and at the same time. We will look hack on that as an anachronism.

How will companies get in the game?

We did it by acquiring technology a little bit ahead of anybody else before people fully understood the nature of what was going on. After we raised $11 million-and our banks had to twist arms to do that-during our IPO, we took all that money and we bet it on telecommunications. If that first acquisition-of modulator technology that gave us a component to enter the WDM business-hadn’t played out, we would have been broke.

You spent $205 million, $65 million in cash and the rest in equity to acquire six different businesses. Now you’re merging with JDS-Fitel; what’s your M&A strategy?

We went on a technology buying spree and bought about $1 billion worth of R&D. We did that because we could see the time to market was going to be key and we didn’t have time for R&D. So we bought technologies that we were betting would be big two years down the road.

The merger with JDS was simple. We had built a portfolio of technology that made us the leader in the “actives” or semi-conductor side of the business, which is how you generate and power the light , in the networks. How you manipulate that light is another set of technology called the passives. JDS was the No. 1 player on that part of the game and we were the player on the other part of the game. So we simply decided that we wanted to be the No. 1 player faster than anybody else. To do that we were going to have to combine the companies.

Each firm has half of the same solution?

We were literally the yin and yang of the business. So we had the same customer base. But to be the dominant player you’ve got to be there earliest with the best technology and market it aggressively. No. 2 doesn’t count. Frankly, who remembers AMD in comparison with Intel? We’ve put both companies on the line. We decided to openly declare to the Street that we were going to have $2.5 billion of good will, which from an accounting sense, meant we wouldn’t make a profit for five years. Turns out, of course, the companies are profitable, generate a lot of cash, have no debt, and are growing very rapidly.

But at four o’clock in January on the Thursday when the press release went out, I thought our companies could be broke by 4:05 simply because the Street would say, “$2.5 billion dollars of good will? We’re going to crash the stocks and you guys are out of business.” We had no idea what would happen.

What happened?

The stocks went up 30 percent. The Street ignored the $2.5 billion dollars. I’m sure if it were $2.5 million they wouldn’t have ignored it. But it was such an outrageous number for these two small companies. In fact, it took me a month prior to pluck up the courage to do it and it was only when the CFO of the Canadian company, Zita Cobb, called me a wuss and I got really pissed off. But it flew in the face of all conventional accounting. That was typical of the gambles we’ve taken.

Do you and Straus get along?

Joe and I get on just fine because we have two quite different roles. He is happiest and most comfortable working with the customer and the technology. I tend to be more of a business guy. He doesn’t like having to deal with a lot of the stuff that I’m prepared to put up with like bankers and people like that. We have skills that are complementary and not competitive. We even talked about who would be CEO, and it wasn’t an issue. He could have been CEO, and I could have been COO. It didn’t matter.

What matters is that telecommunications is by definition a global industry. It’s just not possible to be a purely national player in this business anymore. You’ve got to be able to service customers worldwide. We thought about it on both sides for a long time. But eventually you say history dictates you’ve got to be No. 1, so you’ve got to be there early and you’ve got to be able to maximize your returns.

When you’re fully merged will you have an integrated photonics company?

Photonics is simply the engineering of the particle of light, just as electronics is the engineering of the particle of the electron. Photonics is the word used to describe all the various things associated with the management and manipulation of the photon. But, just as you’ve got memory in the micro-processors, so in photonics you’ve got the lasers. We are not yet at the stage where it’s integrated. We have to make all of these technologies come much closer together. In the final analysis, we are a telecommunications company. We have dedicated our knowledge of the photon to bandwidth, but you can apply photonics to more than just bandwidth. It’s the use of the photon. You can apply it to eye surgery.

If this bandwidth will decrease the need for switches, does that mean that money now spent on switches is being wasted?

No, but the big old mainframe is like the big old central office switch today. There is going to be a need for those puppies. But as more and more intelligence gets distributed towards the edge, you will notice telecom equipment companies seeing a decline in their switch sales and an increase in their photonic sales.

Do you represent a threat to companies like Lucent or Cisco?

No. We’re an enabling technology.

What prevents other companies from developing this technology internally?

Again, history. In the ’60s and 70s all the computer companies were vertically integrated. They did their own software, memory, and semi-conductors, and they differentiated their boxes on a very unique set of technology bases. With integrated silicon and the microprocessor, we saw these companies recognize that the capital investments in these technologies is such that if your only market is internal, your unit cost is too high.

Everything that we understand about the computer business today is going to he true in terms of telecommunications equipment in the future. Rather than actually building these technologies inside, people are going to get them through the marketplace-if there is one that can adequately supply them. Our competition is to be able to make it better, cheaper than doing it internally. If we do that, we’ve succeeded. If we don’t do that, we lose.

Where are you now on that scale, your price versus their price?

Well, it’s not about just price. It’s price and performance. You’ve got to do a better product. In many cases, not all, that is what we can do. Our issue is meeting customer demand. The big fear is capacity. If these big companies can’t feel confident putting their systems’ future in the hands of a merchant marketplace, they won’t do it regardless of price performance. In other words, if they don’t feel they can get enough of the stuff, they will manufacture it internally.

Who will be enabled or threatened by this?

The people who will be enabled are twofold. One, conventional equipment companies like Nortel. Two, startups like Ciena. Whatever subsequently happened to Ciena, it’s a company that had no internal capacity at all and was able to get to marketplace fairly quickly. So both older established companies and new ones will gain from the advent of the merchant marketplace. The consumer is clearly the end beneficiary. The losers are the parts of those big older communications companies that make the switches.

You speak as though this is all inevitable. Is there any competing technology?

There is nothing faster than light. Radio, wireless, electronics, copper-none do it. So from a purely scientific viewpoint, by the sheer laws of physics, we win. And the economics are clearly on our side, because we can carry so much more traffic faster than any other methodology. The challenge is that we’re leading edge. How many people had heard about WDM five years ago?

Nobody. Like the silicon business in the early ’70s, we’re working with some esoteric materials with low yields and sometimes an incomplete understanding of the physics involved. The risk is in scaling manufacturing to be able to take advantage of the new technology. It’s going to be a manufacturing game, just like microprocessors.

What’s the chief barrier to entry?

It’s the know-how as well as the intellectual property. There isn’t an infrastructure in our business. In the silicon business you can buy all the equipment and establish a business. In our case you have to design it, have it manufactured for you, and put up with all the associated prototype problems.

Do your customers accept the concept of unlimited bandwidth?

Yes. There is clear acceptance in the industry. The only question is how do you take advantage of the systems? In our business there’s no one market. We get long distance and metro communications, cable, under sea-and each requires different architectures. The problem our customers face is much more than just the components. They’ve got to write all the software, do all the integration, and choose the architectures.

Lately technology stocks have been hammered, because they are not keeping up with the breathtaking speed. To what degree does this concern you?

I don’t watch my stock price everyday. If I did, I would go crazy because we can have a five- or 10-point swing in a week, sometimes in a day. If you start to run the company through stock price, you will lose. Also, I have no idea why things get valued the way they do. I have no personal investments. I give all my money to other people to manage. I invest my time in running my company. We spend 110 percent of our time worrying about the company, and the technology and the stock price seem to follow.

What are your internal measures?

Every month we do intensive reviews of all of the operations. We do an annual budget. Whenever we’ve done it, we ignore it. We do it ’cause we’ve got a damn board and they want a budget. It means nothing. I’m more concerned about the end consumer than I am about capital investment.

We are changing the way the industries work. Until about 18 months ago, for example, it wasn’t possible to run WDM under the water. We came out with some technology that could actually run this stuff under water. For an increase of about 5 percent in capital cost you could increase capacity forty-fold. Literally within months there was a whole generation of new companies in the submarine telecommunications space-Flag, Oxygen, Southern Cross. A whole group of companies grew up because we changed the economics. It seems silly to think that this little company of ours could do that. But without this technology that we invented, the submarine market and companies like Global Crossing would not be viable. So rather than worry about the state of things at the moment, I’m looking at how we can change economics-the business-in the future.

Microsoft’s rise was almost parallel to IBM‘s decline until it re-configured itself. So what will decline as a result of this?

We’re going to see changes in the older electronic telecommunications technologies, particularly things like central office switches. The rise of Cisco is a clear example of the shift in technology. They compete directly with the switch guys. On the consumer end, it means getting all the eight megabit per second cable, modem, and Internet access into your home, plus video on demand for $100. The choices are coming.

All this depends on wiring the home?

No. Because one of two things will happen. Either the Regional Bell Operating Cos. (RBOCs) will provide Internet to the consumer, or they will be bypassed. AT&T’s goal is clearly to bypass it-use your cable. Three years ago, people would have said it was inconceivable. Two years ago, people would have at least scratched their heads. Now everyone says, “Yeah, it’s possible.” The tidal wave is coming.

What do you need that you cannot readily create yourself?

Well, I would love to have robots, a complete range of automation. I would love to have computer engineering software you can buy in the silicon business. We have to write it all ourselves. There is zero infrastructure for our business. The good news about that is it’s a hell of a barrier to entry. The bad news is it’s a hell of a wall. We’ve probably spent a good half our R&D on manufacturing engineering. What it is or how you get there is half of it. How you manufacture it is the rest.

What is your current bet?

The current bet is the JDS/Uniphase merger. Once we’ve done that and it’s successful, the next big bet will be the new generation of technology. We’re going to need to do much more integration in the semi-conductor level. So, we’ll have to develop new technologies, manufacturing processes, and equipment.

How were you able to see all these things coming together?

This wasn’t rocket science. It was simply seeing what had happened when you get a technology whose time has come. You have to jump on it and bet on it. We eventually bet the company on it. And so there becomes a sense of urgency and an excitement that was unique because you’ve burnt bridges. There is nothing behind you. You’ve got to go forward.

About J.P. Donlon

J.P. Donlon
J.P. Donlon is Editor Emeritus of Chief Executive magazine.