Throwing Down the Gauntlet at ECOC 2019

Avatar photoPublished on
Blog 1569678418 281

On our second blog, it was mentioned that fibeReality tried to get formal confirmation on the coming death of COBO as well as challenged the usual rosy view of silicon photonics at the same workshop. Moreover, there was our disappointing take on the subsequent event in the same conference room. We also had a mixed reaction to Infinera’s XR Optics concept. In this article, we will discuss several other provocations, which took place during Market Focus sessions.

Certainly, the most stunning reaction to one of our questions in Dublin came from Tom Williams, Vice President of Marketing for Acacia Communications, who was discussing the possibility of the marketplace taking too many small steps in achieving 400G without looking to immediately get up to 1.6 terabits in the future. Given that Cisco Systems’ previous acquisition of the DSP provider at 100G, CoreOptics, did not really result in any further developments, we asked Williams whether he was confident that the Cisco-Acacia marriage would lead to the same mentality of leveraging advancements. While there was the perfunctory, quick smile, as the initial reaction, he stated very forcefully that he cannot comment on Cisco, and undeniably wanted to leave the distinct impression that every use case for now was being done independently. The serious and assertive nature of the answer indicated to us that Acacia has to be hurting from the inevitable negative impact of selecting Cisco, rather than say, a Broadcom, to purchase the firm, and that the effectively, former dominant merchant DSP vendor has fully realized that its ultimate fate will be in a dangerous state of limbo for an indeterminate amount of time. Putting it another way, in our opinion, such arguably contrarian statements by Acacia could represent the desperate cry of an unsalvageable situation. Five other notable examples of conflict during the Market Focus sessions include the following:

I. We wanted to know from Mark Nowell, a Fellow at Cisco, whether his remark that the 400GbE is pretty much ready to go and that is has been accomplished quicker than any previous generation of speed was in contradiction to his later comments regarding the pace of SerDes development. He denied it by referring to all of the 8x50G being shown at the show, which the vendor is ostensibly planning to deploy. Nevertheless, after insisting that for the purposes of discussion, putting aside what is being shown on the floor, after a series of quick, follow-up questions by us, he conceded that 4×100-gig would be preferable, including in facilitating 8×100-gig — as well as that it would lead to optimal operational advantages, which many other engineers are pushing, and they would be aspects that would be readily brought to the table.

II. We appreciated the reference to us by Adit Narasimha, Vice President/General Manager, Optoelectronics at Molex that we were first to use the term, “conundrum” associated with Co-Packaged Optics (CPO) in our piece a year ago. Yet, his claim that we employed it in a different context is highly debatable, as he did not rule out the possibility of just a partial use of CPO even at 51.2T. Although fibeReality pointed out that it tends to be a good rule not to blame customers, we said that the unprecedented, continuing tactics in feverishly lowering prices to unhealthy levels, had led to the destruction of the datacom optics ecosystem — and that it is uncertain as to where the funds for developing CPO (not to mention the R&D talent) will be derived, especially if the Web 2.0 firms fail to pony up with investment. He replied that there is the other side to be considered in that these customers have enabled innovation and have pushed for efficiency in penetrating very deep into the supplier chain (we would certainly grant both of these points, at least to some extent). However, Narasimha then mistakenly implied through our inquiry that we viewed CPO as inherently damaging in the same way to the industry as the on-going, vicious cutthroat pricing strategies. All in all, he appeared reluctant to really agree publicly with our indisputable point that the ecosystem collapsed.

III. As opposed to a clash with the individual himself, it was more about interpreting the actual meaning behind Martin Zirngibl’s last presentation as the CTO of Finisar, as he was imminently leaving the company for another one (and that was why he was “in black pants.”) While appearing to be gracious in asserting that he still believed the II-VI deal will result in 1+1=3, it occurred to us that the separate WSS division would account for such a remark. He also took a minor shot at Williams of Acacia, in pointing out that his lines of demarcation between various types of solutions, such as reach versus capacity, were too straight, leading to the false conclusion that everything can be coherent. Finally, Zirngibl made the astute observation that with so much energy increasingly being used, nuclear power should be strongly considered. We then made the point, which he obviously knew, as he stayed silent, that unfortunately, given the politics of the hyperscalers, such a move to that alternative would never occur. While he also did not directly challenge our point on the debilitating actions of the hyperscalers on price, he emphasized that the corporate view was that ownership of fabs is critical to success in the transceiver arena.

IV. Despite Paul Brooks, Optical Transport, Marketing at Viavi Solutions stressing that 400G is a whole new ballgame in terms of a greater amount of efficient measurements, etc., as complexity increases in a market that is heavily going through commoditization, he did not appear to deny our assertion that the best sector to be a player in the optics space is in testing. He pointed out that the new data rate required a step change in rearchitecting the entire testing architecture, and that with all of the moving parts, there is a certain amount of risk in placing bets on R&D expenditures. Furthermore, he talked about the last element for operators to consider are tools for measurement. Of course, with the hyperscalers being such a dominant force, it is not like the old days in which the traditional service providers would wait for a tremendous amount of pain before seriously transitioning to the next data rate (as avoiding the expense of these test systems as long as possible was the rule).

V. The CMO from II-VI, Sanjai Parthasarathi, certainly wanted to remain on message with the silly concepts of “irreversible megatrends” and “transformational forces.” It had to be the reason that he only talked about VCSEL sensors used external to automobiles. So, we asked him: did you skip over the interim use of such devices on the inside of these vehicles as well as airplanes (albeit at smaller units shipped) in the shorter term? So, he mentioned navigation on the steering wheel, monitoring eye balls, and motion traction. In addition, given that life sciences is now a new business category for II-VI, we wanted to know to what extent the FDA and other health regulatory bodies would hold up certain solutions. Although he said such diagnostic/detection systems would not take the same length of time as pharmaceuticals, he did offer four or five years on the former, which is in agreement with our understanding. We think that while there is a tendency to avoid commoditization with such gear for longer compared with other optics spaces, the trick will be in dealing with impatient, public investors on the front-end.

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]