Inphi: De Facto Merchant Coherent DSP Play

Avatar photoPublished on
Blog 1564349571 273

While the bulk of the analyst community in the optics space continues to engage in pandering on the Cisco Systems purchase of Acacia Communications, rather than doing the requisite amount of homework, they also miss out on the bigger picture concerning new opportunities, as well as failing to close the door on other so-called, business prospects. In the same way that the end-users will prohibit Acacia’s DSPs being used in non-Cisco system gear, they will certainly also disallow Ciena and other vertically integrated suppliers (assuming that these vendors are even truly inclined to provide them anyway) from selling such components to their competitors. So, other than the potential for direct sales from the large system houses of components to Web 2.0 operators, which we do not necessarily buy in any kind of big way (other than potentially from Cisco), much of this hype is unjustified (most obviously, the ones running in concert with the PIC-based gear from Infinera) – and appealing margins on the 400ZR will become increasingly difficult. More critically, there is insufficient attention on which player becomes the next Acacia, now that its independence has been taken away. Inphi clearly has the inside track, and we have gathered intelligence that ADVA Optical Networking in Ottawa has partnered with the chip vendor on DSP R&D (and so, we still contend that in-house production is not a pre-requisite for survival).

In the past, despite its acquisition of ClariPhy, Inphi was not considered a vendor, which excelled at coherent technology. Yet, given the prior, extreme, hardball tactics by Acacia, ADVA was looking for alternatives well before the Cisco-Acacia deal (on the last quarterly earnings call, ADVA said, “we’ve been moving to balance that dependency for the better part of 18 months.” So, it is very possible that ADVA heavily encouraged Inphi to get more of a head-start on new introductions — the chip firm mentioned “400-gig ZR modules…going into production in the back half of ’20” at the Bank of America Merrill Lynch Global Technology Conference last month. (In addition, at the Morgan Stanley Technology, Media & Telecom Conference in February 2019, Inphi talked about merging the ClariPhy DSP team with its internal DSP team.)

Naturally, Inphi can count on funding from other system suppliers, frantically looking for other options. Moreover, although it has appeared to us that Inphi was recently giving signals that it is looking to exit, the transaction involving Acacia now provides a substantial chance to move beyond its impressive achievements in PAM4, to counter the negative commoditization aspects of 400G-ZR, and to really dominate in a monopolistic type of situation with the rest of the merchant coherent DSP space.

There is the other advantage that instead of becoming trapped into (similar to what happened to MACOM Technologies Solutions) the “hyperscaler vortex”, with the COLORZ concept, it will now be closer to an even playing ground in negotiating with the cloud players. Inphi also essentially acknowledged at the “BoA” event (referred to above) that the “require[ment for] dispersion control” was looked upon as a negative aspect with COLORZ, as we earlier discussed “trouble with the link…budget,” and there can be little to no doubt now, that our assertion, close to a year ago, that it was always really a Microsoft-centric solution, turned out to be true. 

Menara Networks may conceivably take on the role of the original “ClariPhy,” in making its entrance, particularly at the high end, further down the road. We believe that the supplier is currently behind on coherent DSP technology, and with the changing competitive landscape, we would not be surprised if its parent company, IPG Photonics, brings on a much larger number of developers. Logically, there is the obstacle of Menara being a totally new player to the space, while Inphi has a well-established record at the lower-end already.

As always, NTT Electronics is in the mix, in terms of developing its own higher-capacity DSP, and it has a well-earned reputation for being very trustworthy. Unfortunately, the same cannot be said totally about one or more of the non-Japanese system customers historically, which despite efforts done in good faith by NEL, another DSP offering would be selected out of expediency.

That said, being part of the incumbent ISP, and with so much of the funding coming from the Japanese government, NEL even trying to position itself as a neutral entity for optical systems is virtually unfeasible. Still, it should be noted that NEL has worked with the ex-Oclaro in the past concerning coherent DSPs.

Regarding the Chinese system suppliers, Huawei Technologies, is apparently in good shape with its vertical integration. Others in the country are anxiously waiting for their own version of a non-aligned, domestic DSP manufacturer.

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]