After recently discussing Acacia Communication’s seemingly compelled retreat to a level further down on the food chain, MACOM Technology Solutions Holdings, which has products that are no higher than that third layer has a tendency to embellish expectations, most notably this month on the future potential of the data communications portion of the optical market. At its Cloud Data Center Forum for Analysts and Investors in Boston, there was an emphasis on “Moore’s law pace of innovation,” “the need to be having…terabit discussions today,” and the requirement “to move to a merchant microelectronic semiconductor model.” In examining MACOM, there is also the more practical consideration in that although it provides just about all of the piece-parts that will go into an optical transceiver/module, much of them were obtained through acquisitions, of perhaps varying quality, and so, there may be still a challenge in fully absorbing all of them in order to maximize margins.
At the OIDA Workshop on Manufacturing Trends for Integrated Photonics this past March, there was an impressive presentation by Frank Levinson, one of the founders of Finisar, in which he asserted that “Moore’s Law Has Ended.” He showed that there has been “no real change” in capability between the mid-2012 MacBook Pro Retina and the 15-inch MacBook Pro Space Gray made available in late 2016. Levinson also stated that “[t]he same slowing is happening for optical communications as evidenced by pushing into Shannon’s Law and having multiple bits per Hz.” In addition, at the OIDA Executive Forum 2017, we recall the executive from Finisar, which is the most datacom-centric of all the large optical component vendors, tempering over-enthusiasm about the outlook for demand for 400G based on the very small number of potential data center buyers.
Regarding the transition to the microelectronics-like model, we can find no public reference to the concept by MACOM until after the Senior Vice President for Technical Infrastructure at Google mentioned it at the Plenary at OFC 2017, and the equipment supplier’s CEO exclaimed, “[T]hat was spot on when we heard that. I said, man, that was our vision. That was the vision guiding us through this whole M&A strategy, was we wanted to be the first uniquely merchant market semiconductor vendor into optical. And I think we’ve succeeded.” At a later J.P. Morgan Global Technology, Media and Telecom Conference, after kind of being challenged by an analyst on such a “profound statement,” the rhetoric from the top executive was somewhat muted using terms, such as “pursuing” and “position[ing].”
As we explained on our separate LinkedIn company blog page: “The…Google [exec] not shockingly stated that “a ubiquitous cloud” requires a “revolution” in high-bandwidth data center optics, and that it is looking for “10x scaling” every two to three years…. Google is pushing the envelope to “transformational step functions,” undoubtedly hoping that if the industry achieves a small fraction of these supposed expectations, it will be happy.”
MACOM generates very impressive gross margins for a componentry supplier. Yet, while the hyperbole on the optical side is going strong, it cannot be ruled out that the most sustaining aspect of its profitability tends to based on its legacy business in analog chips for microwave applications.
[written by Mark Lutkowitz]