Finisar: De Facto Owner of VCSEL Space?

Given that we have learned that Avago Technologies intends to keep only the VCSEL chip fab (in March, we were the first news source to report that the supplier was selling such bare die to other module vendors) in its deal with Foxconn Interconnect Technology, our contention is that there is a good chance that a true duopoly likely no longer exits in the space, and that Finisar may by far become the overwhelming dominant vendor in the space. Simply put, we do not believe that Foxconn’s business model lends itself to becoming a real force in the transceiver business in general. Moreover, its ability to retain the “Avago” legacy, when it comes to 850nm wavelengths that goes back to the mid-1990s with the initial R&D efforts by Hewlett-Packard (later Agilent), will be a tremendous challenge.

There is the caveat that extricating the VCSEL fab from the TOSA and subassemblies may be difficult in the Singapore manufacturing plant. We have also heard a little bit of speculation that this arrangement between Avago and Foxconn may even be terminated. (Curiously, there was no mention of the transaction, first announced in August, during the last Avago earnings call.) Nevertheless, Avago would undoubtedly remain determined to find another imaginative means of becoming just a chip supplier in the transport market. (Of course, during all of these maneuverings, hypothetical or not, Finisar is free to totally concentrate on selling more VCSELs.)

Assuming that the transaction goes through, one would think that Foxconn would consider selling off the transceiver (including the VCSELs) product line to another entity. Perhaps Finisar would be the buyer. If it is another company, Finisar’s significant learning curve advantage with VCSELs, in terms of both development and production, as well as its long-term customer relationships, could ultimately result in a substantial decline in shipments for the acquirer. (It is quite possible that much of the “Avago” talent, especially in 850nm technology, may not ultimately be retained.)

Despite all of the attention provided to the hyperscale content service providers, a reminder is periodically necessary that most of the sales within data centers continue to be at 10 gigabits for use on multimode fiber, commonly at 100 meters. (Interestingly, there has been so much hype in the industry about higher data rates across the board that one investment analyst on the most recent Finisar quarterly earnings call initially seemed incredulous that so much of the vendor’s increase in its sales on the telecom side was for 10G WDM.)

Returning to Foxconn, again with the assumption that everything is moving forward fine with the acquisition, it is in for a rude awakening when it comes to running a VCSEL business, which can get ugly, especially for a company unaccustomed to its characteristics. Even for certain Asian manufacturers, fully embracing a “Walmart” model indefinitely of high volume/low margin without the promising prospects for finding applications in potential applications in adjacent markets in which the customer perception gets too much away from “really cheap,” is not obviously enticing. Regardless of the actual outcome, Avago’s desire to move down the food chain should almost assuredly have the effect of strengthening Finisar’s competitive position in the VCSEL sector.

[written by Mark Lutkowitz]