Acacia Will Likely Push 400G DSP

Avatar photoPublished on
blog 1486467890 104

The message that the Street may be going way overboard with Acacia Communications’ valuation, let alone its new secondary offering, may be beginning to resonate. For example, one can read, “Acacia: A Company That Will Need To Be Resistant To The Ire Of Investors.” Yet, the money to be made by at least certain institutional investors with the price going up and down while still remaining above $100 a share, was obviously deemed to be insufficient. We assume that just as with the first offering, a narrative will need to be constructed in order to support the second issuing of stock, and we suspect it may be that the company will be well-positioned with funding to support and potentially dominate the 400G coherent Digital Signal Processor (DSP) space, just as it had done at the 100G rate.

While some 200G applications are emerging, suppliers such as NTT Electronics (NEL) and ClairPhy will be looking to take share away from Acacia at coherent 400G. In addition, there is a good possibility for in-house designs at that higher speed by systems houses, such as Ciena and Nokia.

At up to 80 kilometers, cloud operators would like low-cost solutions and coherent technology is viewed as too expensive. Still, there is a good chance that other solutions will not be much cheaper, and the Acacia approach may gain traction.

Although the 400G standard for 10-kilometer reaches and below will not be ratified until December 2017, pre-standard hardware is starting to appear. Nonetheless, we would not expect Acacia to enter the 10-kilometer space because of the cutthroat pricing competition there, not to mention coherent is overkill in this sector.

If a company is very efficient, the development of a new, higher-bandwidth DSP costs at least $50 million. Usually, the total gets up to $100 million. It explains why in the past start-ups with venture capital funding have developed the DSP, then tend to sell themselves, and the process starts all over again. Interestingly, Cisco Systems has evidently stopped its program of developing a 200G/400G DSP. (Of course, Cisco acquired CoreOptics for its 100G DSP.)

Apparently, Acacia has been in continual development of DSPs at the higher end. Given the windfall from the IPO and potentially from the secondary offering, there is no question that it would be able to sustain this effort indefinitely and an acquisition would hardly be necessary.

A buyout of Acacia might be more likely way down the road because the market for 400G beyond the three largest hyperscale users, Google, Microsoft, and Facebook will be very limited for quite a long time. To give some perspective, if anybody called Verizon or AT&T and asked to get a 40G or 100G private line, it would likely be met with either amusement or bafflement. Nevertheless, the actual timing of true volume production would not even come up in the investment community because the tale of the bandwidth glut never goes away.

Given the amount of time before the 400-gig market takes off, it is quite possible that a newcomer could come along and develop a product later that is superior to the device that would be offered by Acacia. However, it would at least be somewhat problematic to displace a supplier that is so well-established with these processors at the higher data rates.

In the meantime, fibeReality maintains its shock at both Acacia’s market cap as well as the amount of gall it took for both the supplier and the institutional investors to go for a secondary offering. We still do not believe that its silicon photonics provides a meaningful edge. Those observers of the marketplace who point to the supposed advantage of having all of that integration capability in-house, might want to ask to what extent is Acacia farming out for components not only on the photonics side, but in general.

Please find details on our report: Clash of Optical Component Vendors & Technologies in Data Center Networks. For more frequent updates, please click here.

[written by Mark Lutkowitz]



Comments are closed.