Inphi’s 100G DSP: Another Blow to Acacia

Published on
Blog 1506690131 173

As usual, when a vendor, such as Inphi, makes a series of announcements on products or on R&D direction, the bulk of the attention is always on the most cutting-edge feature or solution, not necessarily the most practical, which would frequently be more inclined to ultimately make the most money for the company, especially over the very long term. Its new COLORZ-Lite is really just an extension of a niche product. There is a lot of emphasis on Inphi’s development of 50-gig and 100-gig IOs, not taking into account how much further 25G should remain competitive, including our expectation that NG-PON2 will likely become the de facto access standard for most of the world (which we addressed in real time during a 5G workshop at ECOC 2017), not to mention the technical challenges in moving to these higher serial rates. Naturally, even with the LightSpeed-III M200, the focus on the name is on 200G rather than the much bigger impact it will have on the substantially larger 100G space (particularly in discussing the metro) in finally offering to the merchant market a DSP product that will go head to head at that speed with Acacia Communications.

We believe that Inphi was wise to stay at the lower end for now on DSPs, targeting the public networks, and avoid engaging with Acacia more directly in the tumultuous DCI space, which could have led to a severe pricing war. While we have been outspoken against the hype over metro 100G deployment, there will obviously be gradual growth in this sector over an exceptionally prolonged period of time.

In our opinion, Acacia’s move to even higher capacity DSPs will continue to be insufficient in business size to overcome the following five factors: 1) the reasonableness in assuming that no other optical component vendor comes close to having so many of its devices stockpiled, 2) even if there is a revival of the China market, it will hardly resemble the previous boom, which was based on an extraordinary amount of pent-up demand; 3) its noticeable failure to come up with an adequate encore to its initial success; 4) resorting to selling its PICs, and 5) the loss of a certain amount of credibility with both customers and investors.

Interestingly, when we talked to Inphi at ECOC, it seemed to go out of its way to mention the technological prowess of Acacia. It is hardly inconceivable that the valuation of the latter will go down substantially, permitting the former to pick it up at a decent price – or at least lead to a merger. Inphi could lock up the DSP merchant market for itself (with the exception of the what we believe to be the nonsensical play by Ciena, not its partners), as well as avoid having to develop its own processors, which would operate at above 200G. Our understanding is that NEL is not much of a threat to either major DSP competitor, given an inadequate amount of funding available for this purpose.

There appears to be signs that Acacia would welcome such an exit. There were apparently clear indications of its future in the fall of last year as expressed in a Seeking Alpha article at the time: “Now the bad news is that shareholders are willing to dump shares nearly $30 from the recent highs and on top of blow out numbers. These insiders see trouble over the next year.” Another post at the same site provided a similar type of warning. While fibeReality does not recommend stocks, we certainly thought that Acacia’s valuation earlier reached a point that was extremely out of proportion with the rest of the market.

There is other evidence that Acacia may be looking to get out. The bulk of the employment opportunities at Acacia seem to be more about bringing on engineers for troubleshooting purposes, rather than product development. Perhaps it was not a coincidence that Acacia had no booth at the big European optical show.

To follow us on our separate, daily company blog, which is exclusively on fibeReality’s LinkedIn page, please click here. Please find details on our “Clash” market reports here.

[written by Mark Lutkowitz]



Your email address will not be published. Required fields are marked *