fibeReality believes that the factor that makes Inphi the most attractive as a buyout candidate is its recent purchase of ClariPhy Communications. With PAM4 technology yet to be sold in any kind of volume by any vendor, Inphi can potentially provide the necessary engineering and financial resources to fully develop a high-end coherent DSP to the merchant market that would provide an alternative to Acacia Communications, which has been able to monopolize the space. As we have discussed in the past, the lack of a second supplier makes differentiation highly problematic for those system vendors, which do not internally develop their own DSPs for these applications. Assuming that NEL does not come out with its own version of this kind of processor, which would be a compelling competitive solution, a duopoly situation would also help ensure that any kind of pricing wars would be kept to a minimum.
A major reason that Inphi cannot offer too much of a me-too device is the speculation in the industry that system suppliers have been doing a certain amount of stockpiling of Acacia’s DSPs. Such an assumption appears to be quite logical with the recent leap in demand for high-capacity components in China in combination with the relatively large number of vendors dependent on Acacia’s DSPs for their Data Center Interconnect (DCI) pizza boxes. Any large customer would also probably want to be assured that before it chose a particular supplier that the manufacturer had adequate access to components including for any unexpected increases in demand. It will be interesting to see to what extent, if at all, Acacia’s future revenue is negatively affected by what would again be rational actions by its customers.
For all of the talk in the industry about PAM4 replacing NRZ, the most prominent future buyer of the new modulation technique, Microsoft, has seemed to be quite deliberate about pulling the trigger on its extensive use. It has not helped that apparently more than nine months ago, we understand there was a budget cut at the large data center operator, which halted purchases, likely up to 40 kilometers in length. Around that time, Microsoft apparently issued a DCI RFI to make sure at least indirectly, it made the right decision in choosing Inphi. According to our recent metro 100G report, “it would be reasonable to expect a second RFI to see how prices have changed in the last several months – that would be typical behavior for Microsoft.” Regardless, we are unaware of anything getting in the way of Inphi’s announced shipments to Microsoft later this year.
It is quite possible that Acacia itself may want to merge with Inphi, resulting in a combination, which would possibly be even more compelling than the two entities individually for a potential acquisition. On the other hand, several of the much bigger semiconductor firms would probably be interested in purchasing Inphi itself.
With our continuing expectation that Huawei Technologies will get into the optical components business in a meaningful way, it could not be ruled out that it would offer its own coherent DSP that would be competitive with both Acacia and Inphi. While we would anticipate that it would not even come close to the feature/functionality of the DSPs that would be part of its own full systems, it might be a scaled-down version, which could be similar to the devices from the two major competitors, but obviously priced very aggressively.
Please consider another one of our latest reports, Clash of Optical Component Vendors & Technologies in Data Center Networks.
For more frequent updates, please follow us here.
[written by Mark Lutkowitz]