The new millennium will usher in business practices and technologies that promise a highly scalable marketplace that will be able to support the burgeoning demand for high-capacity optical-communication systems.
We have spoken often in this column about the high rate of growth in the optical-communications market. With those discussions have come periodic discourses on a variety of shortages. At one point, it was 980-nm pump modules, and on another, thin-film filter-based wavelength-division multiplexers. Those topics have led us down the path of questioning the health of an industry that seems to have problems scaling.
As we exited 1999, many of the signposts of a change in this condition had begun to fall into place. For the optical-component industry, the new millennium will usher in business practices and technologies that promise a highly scalable marketplace that will be able to support the burgeoning demand for high-capacity optical-communication systems.
In recent weeks, we initiated coverage of Newport Corp. with a buy rating. A relatively small portion of Newport's revenue comes from automation-infrastructure products sold to the optical-component world, but our thought was that we may be on the cusp of a change in that regard. Heretofore, the optical-component business has been a people-intensive effort that would boggle the mind of any chief operating officer in the semiconductor industry. There has been no such thing as economies of scale. One high-ranking industry executive we spoke to about the subject quipped, "If we continue on our current path, we will need 45,000 employees and 8 million square feet of manufacturing capacity to meet our customer's expected demand over the next three years." The company isn't a tenth of those figures today.
So what's going to change? Well to start with, the mindset of the companies will change. One example was the late 1999 appointment of executives specifically responsible for process automation at JDS Uniphase and SDL. Both executives are out of the semiconductor industry. Supply issues have become a top priority at both companies, and we believe they are also high on the radar screen at many of the other companies in the optical-component sector.
The evidence is widespread.
- JDS Uniphase purchases Optical Coating Laboratory and SIFAM Ltd., reportedly to ensure supply of critical components.
- E-TEK Dynamics invests in capacity in Taiwan and mainland China to gain access to lower-cost labor.
- E-TEK funds thin-film-filter capacity at a number of its outside suppliers.
- SDL talks of more than tripling 980-nm pump-module capacity this year.
- Corning makes plans and a major acquisition to more than double internally supplied components.
We would go so far as to say that in a market where the competitive banter can get pretty heated, détente seems to have broken out in the last few months. There is talk about plenty of business to go around and about greater cooperation on efforts to reach standards in the packaging arena.
All of this could set the stage for a production-equipment environment that can move from custom-made to merchant-supplied systems. Veeco seems to have picked up the scent as evidenced by its purchase of Ion Tech, a small supplier of ion-beam deposition systems used in the manufacturing of thin-film filters. We are seeing other evidence of the traditional semiconductor capital-equipment players turning their attention to the challenges of the optical-component market.
Will these products perform well in high-volume markets? That remains to be seen, but we had the good fortune while doing some background work on Newport to speak with some customers. One executive, formerly of the semiconductor industry, suggested that we recall the beginnings of that industry. He pointed out how people questioned whether Applied Materials and others had what it would take to displace internal efforts as well.
Where there is a need, there is an opportunity for focused, forward-looking companies to capitalize and do something people don't expect of them. Harmonic emerged from a year-and-a-half of lethargy in late 1998, with few investors who believed in them, and became a star performer supporting the upgrade of cable networks to multiservice-communications networks.
We can't say with certainty that Newport or Veeco will emerge as leading merchant suppliers of infrastructure to the optical-component market. What we can say is that they are on a short list of companies that are taking a run at the opportunity. That gives them a chance.
Automation, greater industry cooperation, the adoption of better industry practices like EDI, and other tighter ties between customers and suppliers promise improved industry scalability over the next couple of years, but new technologies have even greater long-term promise.
It is unlikely that a new entrant to this market will simply duplicate what today's incumbents do. Unlike the semiconductor industry, there are no suppliers of standardized equipment and software that permit new entrants to quickly build capacity; much of the manufacturing hardware and processes are internally developed. As a result, this situation presents a barrier to entry into this market.
Instead, when looking for potential new investment opportunities in the photonics-component marketplace, we feel that the best bets will be in new or evolving technologies that in some way address the supply constraints.
While not a new technology, several young companies, including ITF Optical and Wavesplitter, have been developing processes to mass manufacture WDM devices based upon fused biconic taper (FBT) technology. Devices employing this technology benefit from highly efficient transmission due to the fact that these are all-fiber devices. That is, the light never has to leave the optical fiber, so no coupling is required through multiple components or into other waveguides.
The all-fiber geometry also has manufacturing benefits. It enables an interesting manufacturing model. Whereas multiplexing and demultiplexing components based upon thin-film filters requires having a supply of a large array of thin-film filters--essentially one for every wavelength--FBT multiplexer/demultiplexer devices have no such issue. The performance features of these devices are obtained purely through the manufacturing process itself. That is, no additional or specific components are needed to go from manufacturing a 16-channel device to a 40-channel device. The only material needed is the fiber itself. The performance features are obtained by splicing, fusing, and writing gratings into the devices. This approach permits a "manufacturer-to-order" model with minimum requirements on inventory, which should prove beneficial to an industry facing severe supply shortages. These facts enable FBT technologies to scale in channel counts efficiently as well as meet customized network specifications without significant capital expenditures.
Similarly, arrayed waveguide gratings (AWGs), while not new, are likely to grow in importance. Once again, the ability of AWGs to function as DWDM multiplexer/demultiplexer devices comes not from the use of multiple components but from the process of writing multiple waveguides onto a single substrate. Adding channels requires little more than writing more waveguides. Manufacturing AWGs employs lithographic methods similar to those used in the semiconductor industry, and they are therefore highly scalable devices. As systems channel counts grow in number, AWGs are likely to grow in importance. Unlike FBT devices, however, coupling of light into and out of the waveguide remains a manufacturing issue and potential bottleneck.
Among the most unique set of technologies that present a solution to the current supply shortage is that being promised by Nanovation Technologies. Nanovation is proposing to develop integrated DWDM components by writing optical waveguides onto planar semiconductor substrates utilizing proprietary lithographic processes. Since these processes are based upon those in use in the semiconductor industry for the mass manufacture of electronic processors, Nanovation's technology has the potential to permit mass manufacture of planar photonic devices. Such an ability has been crucial in the development of the semiconductor industry and could prove equally so to the photonics industry. Nanovation's devices measure tens to hundreds of microns on a side. As a result, there is the potential for thousands of devices to be manufactured from a single wafer.
Clearly there is a need for this level of scalability in the photonics-component marketplace, and in a relatively short period of time, the incumbent suppliers and emerging companies appear to be tackling the challenge despite what could be considered overwhelming demand.Robert Mandra is an associate in investment banking with SoundView Technology Group (Stamford, CT). Previously, he was an optical engineer with MIT Lincoln Laboratory for nine years. He can be reached at (203) 462-7361 or [email protected].