The optical industry was standing by for Acacia Communications’ S-1 for a lengthy amount of time, and there had been some recent conjecturing that it may not happen at all. The stakes were clearly high in that it appears to have been the only VC-backed 100G firm. While Acacia has made impressive penetration into the market with superior DSP technology, the question surrounds what it can do for an encore that would be good enough to attract investors in a meaningful way on Wall Street.
Certainly, the process in Acacia going public was hardly done with a lot of fanfare, reflecting the current relatively low number of IPOs occurring in the technology sector in general, and the close to non-existent activity in the fiber optic space. The vendor had filed confidentially back in August (actually a common practice now for all companies with yearly revenues of less than a $1 billion). Then, very close to the holiday season (which was probably not a coincidence), Acacia noiselessly notified the SEC of its intention to go public around the start of 2016.
Even competitors would have to acknowledge that Acacia has the best DSP in the marketplace, especially in terms of its low-power dissipation. It did not hurt that the vendor introduced the solution early in the game. In addition, Acacia’s CFP excelled at facilitating 100G coherent capability, allowing it to be accepted at various system vendors, which did not become vertically integrated at that data rate, including ADVA Optical Networking,
From an actual timing perspective, a transceiver manufacturer going IPO now is not too bad in that major players, including Lumentum Operations, Oclaro, and Finisar are having difficulty keeping up with demand. We realize that a big reason for the increase on the telecom side for these components is that large service providers in China have been transitioning from 40G to 100G. Of course, this expansion effort will eventually start to flatten out.
Another aspect to consider is that all of the current DCI rhetoric will help Acacia. A good number of the 1RU boxes seem to be constructed around its 200G module. Nevertheless, we have recently addressed the fact that only about three of the hyperscale operators have a present requirement for such capacity.
Concerning the need for more of a sexy, potentially high-growth story for the Street, Acacia’s “Silicon Photonics” (“SPs”) transceivers will undoubtedly be heavily promoted. While they are evidently designed quite well, (and as Intel is still striving to save face by extensively promoting the viability of its own “SPs”), the optics on Acacia’s devices apparently do not set them apart from other componentry in the business, other than to be additional window dressing for the DSPs. (At least in the past, Acacia implied that one would see integrated photonics at 400G instead of 100G – now it is not really making much of a distinction between integration and “SPs.”
At the end of the day, it still comes down to Acacia’s request for additional capital based on the vague requirements in its S1. In lacking a CFP-ACO, the company has begun selling its DSPs externally, not only weakening its revenue model, but leading to concerns about the actual potential of that business for processors. So, something like spinning a new 16nm chip would probably not lead to a tremendous amount of excitement with public investors. (There is also the constant challenge, especially with a smaller firm, regarding the ability to deliver a sufficient number of modules to its customers.)
[written by Mark Lutkowitz]