Makers of Gigabit Ethernet transceivers find the market only gets tougher

By Mark Lutkowitz, Trans-Formation Inc.

Table of Contents
Driving the market with volume
Coming market explosion
Limited earnings potential
Investor, be discriminating

In the current climate of fiber optic fever, there seems to be a sense that any company that is a player with this technology is a clear winner. While not every new optical initial public offering (IPO) has been assessed with equal weight in recent months, at least some fiber optic companies have certainly been given initial valuations that could be viewed as excessive. In several cases, there appears to be ignorance of the historical market dynamics of specific sectors of the fiber optics industry. One example of this trend is in the Gigabit Ethernet (GE) optical transceiver space.

The difficulty in effectively competing in the GE transceiver market was evident even two years ago. Suppliers were complaining about price deterioration and how fast the margins were being driven out. One executive at a large transceiver company claimed at the time that if the commitment had not already been made to the space, the supplier would not be in the business. Price constraints appeared to have been escalating in the Ethernet transceiver space since that time.

Driving the market with volume
Historically, prices of these transceivers declined partially because of Hewlett Packard's (now Agilent; Palo Alto, California) general philosophy of trying to drive volume and respond to the marketplace with the lowest prices of any supplier. Another major reason was that for the last several years, Ethernet system suppliers have been pushing the model that Intel (Santa Clara, California) has always used for processors. All of the big enterprise system vendors want to mark down their switches every month or so. Consequently, they need to lower their costs to keep their margins up, and so a lot of pressure is put on the transceiver manufacturers to decrease their prices.

Despite this history of being a very tough market in which to make substantial earnings, Methode Electronics (Chicago, Illinois) successfully spun off its optical components business that had a heavy focus on optical transceivers. The new entity, Stratos Lightwave (Chicago), was able to take advantage of investors' recognition of the previous success of one its major competitors in going public, Finisar (Sunnyvale, California). While Methode's optical business was making a modest profit at the time of the IPO, it was publicly known months before that it had not had a positive cash flow (see Table).

Coming market explosion
Besides the euphoria over optical IPOs in general, the other factor that clouds the issue with optical transceiver valuations is the impending explosion in the GE system market. Without a doubt, unit shipments of transceiver modules will dramatically escalate over the next few years as the GE system vendors make huge volume purchases of these components. It will probably be at least three years before users are going to exhaust their GE requirements and need the next generation 10 GE speed. However, the margins on these one GE transceivers will remain small.

Ample evidence of this reality can be found by the actions of another leading supplier of GE optical transceivers, Finisar. Although Finisar is expecting sales of its transceivers to climb rapidly, the supplier has been very aggressively using its large amounts of available cash to diversify higher up the food chain from the component level to subsystems where the profit margins are inherently better. Stratos itself is making moves into the telecom component space where the margins look more attractive.

There are several reasons for the GE optical transceiver market continuing to get worse and not better in terms of profitability. First of all, the transceiver simply sticks out like a sore thumb. Although the component is just an optical-to-electrical interface, it makes up a significant cost of a GE system. In fact, it is usually the highest cost element of the system, and so it gets the most attention. As much as 20% of the actual system cost is made up of this optical module.

In contrast, a dozen years ago the cost of interconnection was just 5% of the cost of an enterprise system. So, when system suppliers are trying to reduce costs, the transceiver vendor is going to get caught in the crossfire. Furthermore, as the industry starts working on 10-Gigabit Ethernet, there will be even more demands to get 1-Gbit/s components driven down in cost.

Limited earnings potential
Another big reason for paltry earnings potential is the lack of any meaningful opportunities to differentiate the product. Right from the beginning, the goal in the industry was to make GE just like any other interface with quick standardization. Consequently, the only alternative has been to engage in price wars with direct competitors.

Moreover, although transceivers look cheap from appearance, they are fairly complex and costly to develop and manufacture. A key to long-term success for any transceiver supplier is based on its investment in at least some vertical integration. The ultimate choice on vendors by the system suppliers will tend to be based on the optical module suppliers' ability to produce as many of the individual electronic and optical components of the transceiver as possible. By controlling most if not all of its processes, the optical module vendor is not victim to shortages from outside suppliers and is also in a better position to control its costs.

Lastly, there will be price pressure on these component vendors caused by the sheer large number of new optical transceiver suppliers being funded. Venture capitalists are throwing money at this segment without much discretion and with little understanding of such a low margin business even though companies have and will continue to exit this market, and even though they have to compete against well-established suppliers including IBM. Probably only a relatively small percentage of these new companies are going to intelligently invest in the manufacturing infrastructure necessary to be really successful long term.

Investor, be discriminating
All in all, it is clear that not every fiber optic opportunity should be even close to being evaluated in the same way. Still, it is just human nature to lump disparate opportunities under a single category and look at the whole space superficially because loads of money are being made in general there.

There is some evidence that not too long ago Methode was beginning to slow down its market penetration in fiber optics because its cash cow was really its power business. Yet, with its responsibility to its shareholders, it is hard to blame Methode for taking advantage of the current environment to rake in some big bucks with a business that is hard-pressed to generate profits in its current position. Stratos Lightwave has a lot cash in its war chest now to make further investments to better gear up for a low-margin business as well as to further diversify its product line.

About the author…
Mark Lutkowitz is president and principal analyst of Trans-Formation Inc., based in Birmingham, Alabama.