Lumentum: Six-Inch VCSEL Retrospective

Avatar photoPublished on
Blog 1542403140 245

In view of fibeReality’s latest analyses which demonstrate the essential role that VCSEL wafers are planned to play in future product designs of all kinds at Apple, along with both of Finisar’s last two CEOs hardly appearing to do a satisfactory job in execution (although the vendor did have some hard luck on occurrences outside of its control), regarding their desires to play a dominant role in providing them, several of Lumentum’s public statements in this space can now be legitimately deemed as at least, arguably rational. In our opinion, whether they were all ethical and defensible, is a whole other matter. Certainly, before anything was actually shipped, Lumentum’s talk about a minimum of 50% gross margins, which could easily be characterized as a rapacious action, and would have been otherwise considered to be close to a suicidal type of move regarding its relationship with Apple, at least, debatably has a minimal context. Also, the almost immediate move in its messaging to the supposed large potential use of Edge Emitters (EEMs) in smartphones, could now be considered logical, but notably only from a shallow public/investor relations perspective. (Although a similar commitment from II-VI may have been a red flag, before its announced purchase of Finisar, we just did not perceive the former as necessarily having a powerful ambition to be in this space, independently anyway, despite its investments in the Anadigics and UK fabs). In addition, the discussion of Lumentum utilizing Oclaro’s fabs to produce VCSEL bare die can now hopefully be taken with a grain of salt.

We now know that Lumentum was Apple’s second choice at best from the beginning. Moreover, there can be little doubt that the optical supplier had to be as stunned as Finisar that WIN Semiconductors came through with volume shipments of six-inch dies just in time. (Yet, there was not even a hint expressed in advance by Lumentum about the apparent chaos over technological issues during the summer of 2017.)

Thus, Lumentum’s leadership may have decided to totally milk what it could have viewed as a short-term opportunity to get its valuation as far up as possible, and strive to maintain it at a high level. If nothing else, it allowed the firm to very comfortably go after Oclaro, when it desperately needed to make such a move, given the breakdown of the data communications optical ecosystem.

Lumentum, as well, could not have expected Finisar to play a game of chicken with Apple in order to try to get the latter to put up the funds for the new manufacturing plant. Once again, the several-month stoppage of negotiations allowed Lumentum to get in front on the account.

Of course, it is not as if Apple was trying to keep its preference for Finisar a secret for long, and the former provided that infamous dog-and-pony show. Yet, Lumentum could not have anticipated that technical issues, at least initially, really only allowed for a single vendor, and that the presumed volumes for these chips would wind up being too optimistic.

Furthermore, Lumentum likely would not have believed Finisar to be so stubbornly persistent about the partial viability of four-inch devices in smartphones. Actually, we would not be surprised if there are still some managers at the latter, almost assuredly not directly in VCSELs, who still believe that the smaller gear could be substantially used by Apple. Most shocking to Lumentum must have been Finisar’s insistence on not making minor alterations to its process order at the Sherman plant.

On a side technical note, we definitely continue to have trouble seeing an application for EEMs in smartphones. For a short distance application, VCSEL arrays have a big advantage for Time of Flight (ToF). With a VCSEL, one can place a diffuser chip in front of the array, and illuminate a large area, and accomplish it with a nanosecond pulse.

Once more, the requirements of size, and especially depth (along with wall-plug efficiency) are different with a VCSEL compared to an EEM. It would not matter if a thermal electric cooler was used because such things would not be put in a phone application. For a LiDAR app, in which one wants to make a small dot far away, then EEMs are more beneficial for ToF because of their higher brightness.

Returning to the theme of the article, it is still quite possible we are being overly generous with this exercise in attempting to give Lumentum’s CEO the benefit of the doubt (albeit, not a justification), chiefly involving the exposure of expected margins from a single customer’s purchases. In our estimation, other buyers will still question his judgement on this matter, as well as perhaps on the past remarks regarding both the future of Oclaro’s fabs, and the use of EEMs in smartphones. There is also often a high price to be paid ultimately in both the investment and user communities for mounting disingenuousness by a corporation, regardless of the motivation.

As always, fibeReality does not recommend any securities, and this writer does not invest in any companies being analyzed by us.

To follow us on our totally separate, quick update company blog, which is exclusively on fibeReality’s LinkedIn page, please click here.

[written by Mark Lutkowitz]