As fibeReality pointed out in a recent article, the counseling from Hock Tan, CEO of Broadcom, to his former executive on the wireless side of his business, Michael Hurlston, who became head of Finisar, was to make an already terrible situation become as unrecoverable as possible. We understand that it had at least the appearance of a tag-team effort, as Ford Tamer, CEO of Inphi, another alum of Broadcom, became an active participant, as he made the offer (which again, II-VI CEO, Chuck Mattera blocked) to purchase Finisar’s IC group. His interest in trying to help Tan further debilitate the large optical module vendor, including in encouraging the foolish change in strategy to a contract manufacturer model, seemed logical, partially in that we expect Inphi to be acquired by the chip giant down the road. Obviously, Inphi would have also benefited in the shorter term by acquiring these ICs at a ridiculously low cost. However, while one should certainly avoid immediately jumping to the conclusion that Broadcom has plans to offer a complete line of optical transceivers, there is also no doubt that its market position with optical chips has become weaker.
After the “divestiture” in 2015 to Foxconn Interconnect Technology (FIT) of these devices, Broadcom remained intimately involved in the space, and so is quite aware of the extraordinary pricing pressures that have occurred, especially in the data communications sector. It seemed to be a totally captured situation by the latter of the former. Seemingly, FIT was not allowed to source components from other suppliers.
We foresee Broadcom tending to take an exceptionally targeted approach, which will potentially lead it to command a premium for these products. We do not think that the additional revenue gained in offering full modules by itself would be the priority, in accordance with Tan’s modus operandi. Yet, given the lack of non-Chinese, dependable players in the space, we would not rule out certain opportunities in the future to go after even conventional applications, resulting in acceptable margin generation.
We would also not be shocked to see Broadcom going back to the old “Avago” model of developing differentiated/proprietary packaging approaches. The idea would be similar to the somewhat exclusive VCSEL-based engines in previous years. Additionally, other than Inphi in the longer term, purchasing one or more vendors, like a Semtech, which enables an InnoLight, could be a possibility.
Regarding chip optics, Broadcom has been making less money because of increased competition, even with datacom VCSELs. Most conspicuously, it was unable to keep up with other vendors on InP technology. Nonetheless, there is apparently a high probability that the firm will start R&D efforts on simple, CW InP lasers.
Moreover, with the new SVP/GM of the Optical Systems Division, Alexis Black Bjorlin, there has been the assumption by some people in the industry that Broadcom will now be moving aggressively in a silicon photonics direction. While in fairness, our latest intelligence indicates that she should receive enormous credit through sheer will to ship a few hundred thousand 100G CWDM4s at Intel, mainly based on the SiPh platform, few people in the world have a better understanding of both the cost challenges in general, and the technical hurdles specifically, in moving to higher rates.
If Broadcom decides to move forward with a SiPh program, it would be the polar opposite of the Intel situation, which was based on a loss-leader model. Tan’s response to inadequate performance would be swift and unmerciful.
For our fast-growing, totally separate, quick update blog, which is exclusively on fibeReality’s LinkedIn page, please follow us here.
As always, fibeReality does not recommend any securities, and this writer does not invest in any companies being analyzed by us.
[written by Mark Lutkowitz]