During the firewall period, waiting for the merger with Finisar to be completely approved, from the CEO of II-VI, Chuck Mattera, people in the industry got the impression that he expected the combined company to become kind of the “Intel” of optoelectronics. Putting aside the fact that the actual Intel has played a prominent role in helping to destroy the data communications optical ecosystem, and perhaps is not the best example to use in that context, II-VI will be getting on a me-too line of large component companies, which have either begun the transition from modules to chips or are involved in contemplating the idea. Broadcom and Lumentum are the most notable players. There is Sumitomo, which is similarly looking to both move down the food chain, and we would not be surprised if it decides to offer VCSEL bare die in a big way for data communications applications. In addition to our previous discussion on Finisar, pre-closing, with our intelligence indicating that it had abandoned customer support on datacom VCSELs, a lot of people in the rest of the transceiver (TX) business there had already either jumped ship or tried to get out of the TX group in anticipation of moving downstream in joining II-VI. Then if one adds all of the specialized chip vendors, including Inphi, Semtech, etc., which may ultimately become part of bigger combinations, we do not envision anything necessarily special about II-VI’s plan, or the supplier as a whole itself, as it has inherently not tended to distinguish itself historically as a top-tier photonics vendor, and it will not help that the distinctive Finisar culture is expected to disintegrate in the same way it is happening to the ex-Oclaro individuals at Lumentum. Moreover, we continue to have questions over the judgement in general of the leadership at II-VI. Finally, there needs to be consideration of the interesting dichotomy with the industry shakeup taking place between optical system vendors looking to increasingly become vertically integrated and more proprietary, including Ciena, Infinera, Cisco Systems, and Infinera (specifically with the Groove platform), while the hyperscalers continue to push for openness as well as uniform piece-parts.
While Mattera directly witnessed what happened to Agere/Lucent Microelectronics, clearly, the market dynamics were different at the time with the bubble at the turn of the century, as well as being divisions of system houses, and so, the disappearance of such large component entities in the traditional marketplace may not happen, again. Of course, during those days, there was also far less emphasis by dominant operators in making a concerted effort to take optics equipment companies out of the game like there is today.
However, while several of the bigger optical system players, are continuing to support a disaggregated approach with their rhetoric to appease the hyperscalers, they are actually increasingly refusing to play the game of not being in as full control of the supply chain as possible, even though some of it may be external to the company, as in the case of ADVA Optical Networking. Regardless of the higher margin generation in moving to the chip level, large datacom/telecom component companies, lI-VI and Lumentum may somewhat wind up in the undesirable middle between the conflicting goals of the major parties on either end of the food chain.
Concerning fibeReality’s opinion of the wisdom of very top-level managers at II-VI, we cite the following eight examples: 1) purchase of the Warren, New Jersey facility, at least originally, was not considered to be in the class of a superior production plant; 2) acquisition of the large Newton Aycliffe fab, which has been an albatross around so many necks (maybe partly explaining why an answer was not provided to us on its future at EPIC’s recent conference on VCSELs); 3) concern in the marketplace that its joint GaN development effort with Sumitomo may suffer because of distractions with 3D sensing; 4) making the same mistake (albeit not as specifically egregious) as Lumentum in bragging about not only its future margins going up on 3D sensing, but did so in a laid-back, game-like way, which had to be disconcerting to Apple; 5) Chinese operators insisting that Finisar’s WSS division be kept separate for a few years; 6) over-the-top messaging on its prospects, not only in the language, but totally skipping over the use of in-cabin VCSELs in a presentation of automobile sensors (as we pointed out during an ECOC event); 7) failing to give a CXO position to anybody from Finisar; and 8) (quoting ourselves) “the insensitivity to highlight an executive, who obviously had to accept a demotion.”
Perhaps one needs to look at both II-VI and Lumentum, not just as holding companies, but more accurately as venture capital firms with product portfolios. While diversification in the optics space is a must, the quite nonchalant way that life sciences was announced as a new business category by II-VI at ECOC, given the upfront regulatory challenges (which we brought up), it sounded more like a VC than a componentry supplier. A corporation in the latter category would be truly devoted to very meaningful and quite extensive R&D work, as opposed to the priority being investment in various assets and hoping for one or more to result in an extraordinary return.
As always, fibeReality does not recommend any securities, and this writer does not invest in any companies being analyzed by us.