We have been pondering the question of what happens when emerging optical component vendors, being financed by mega-data center operators to achieve the lowest cost possible, need more money, especially for new research and development purposes. The best answer, which is starting to become self-evident, is for the componentry company to go public, because with the undeniable cachet of a Google-backed firm (in the case of InnoLight Technology, Google Capital), the highest valuation can likely be achieved. The perceived worth of these newcomers in being purchased by a larger competitor in this presently unattractive business would apt to be much lower as there would probably be a considerably higher level of scrutiny given to matters like the potential to sell their devices to other than their “captive” buyers.
For Google and other prominent enterprises, the IPO route for these firms provides the cleanest outcome in that it helps to ensure the following four results: 1) a full ROI; 2) no additional financing or increase in gear pricing 3) a lengthier relationship with a supplier, which has fully customized its solutions to its technical needs; and 4) no vested interest in the long-term market performance of the vendor with complete flexibility to make investments in new solutions that can drive their optics cost even lower.
From the perspective of the component firm, the potential for the IPO option of providing a greater amount of cash will be important in penetrating other accounts. While we would be able to say now that there is a lack of any noticeable prohibition on InnoLight, Kaiam, etc. supplying entities that are not financiers, in effect, the amount of initial funds raised may result in an inability to scale up manufacturing capacity to accommodate others anyway.
Regarding Kaiam specifically, with both Google and Microsoft evidently backing it, we obviously would not be surprised if the IPO exit strategy is adopted once again. Yet, one difficulty that Kaiam (along with InnoLight and potentially others) will have to face is that in keeping costs down in not having to design products to meet the complex requirements of the industry standards, such as very low loss, a lot of potential customers may not be willing to put up with such sacrifices. (As we have mentioned in the past, each of these large end-users in effect has its own set of standards.)
It is also interesting to look at the use of Planar Lightwave Circuits (PLCs) by Kaiam. The apparent problem with a PLC is that it is very long – multiple millimeters, and although the vendor is pulling it off, it is hard to see how the circuit fits into a QSFP28. The bigger concern is whether just the mere appearance of a significant relaxation in performance is going to make it prohibitive to other users.
Of course, the most important consideration in our discussion of potentially indefinite sustainability of the InnoLights is once again, the well-established, much bigger optical component vendors getting caught in the undesirable middle – with the systems integrators on the other extreme also continuing to take hold of the most attractive business opportunities.
[written by Mark Lutkowitz]