Infinera’s Last Ditch: Infinity and Beyond

Avatar photoPublished on
blog 1486467815 87

With the full realization by industry analysts, some of whom initially drank the Kool-Aid, that the potential for Infinera significantly penetrating the metro optical market will not be a pushover, the supplier’s frantic gambit is to prematurely advance a marketing message to the extreme on the Long Haul (LH) side with its Infinite Capacity Engine (ICE). With the bulk of the LH carrier space continuing to migrate to 100G, its Gen 4 technology supposes the potential for well over an order of magnitude increase in capacity deployment. A reasonable presumption is that in the vast majority of cases, it will take a very long time to reach exhaust at 100-gig with the use of DWDM. (Infinera dodged a question about the relatively low number of wavelengths at that speed during the last quarterly earnings conference.) The company’s arguments about a direct correlation between expectations of bandwidth growth (especially involving applications that have not yet arrived) and high-speed equipment purchases are hardly convincing, especially based on historical trends. They also have to be tempered by the comments of the CEO on the earnings call and his acknowledgement of “a little bit of pensiveness around me until I see more results of what’s happening in the broader market,” which would also affect metro sales as well.In our opinion, the future ramifications of the past grandiose outlook for its growth potential that was expressed by Infinera in the recent past concerning the metro space and the expected technological difficulty in just pulling off its 2.4T chip (even putting aside the anticipated conservative increase in bandwidth levels) will likely result in continuing dire consequences for the company.

There was also the the admission on the earnings call that there is no more room to grab additional LH market share in North America. In effect, Infinera is indirectly conceding that there is no potential of getting business from either Verizon or AT&T, by far the two largest service providers in its own backyard, both of which have obviously determined that the supplier is still not considered to be ready for prime time even through an OEM arrangement with incumbent suppliers. In the metro sector, it could easily be assumed that the vendor’s lack of an adequate track record would automatically preclude it as well from consideration. While on the call, Infinera can state, “We’re building a great business with or without the AT&Ts and Verizons,” its inability to convince these customers to buy its products is a blow to the supposed competitive advantages of its PICs expressed throughout the years.

Potential customers will readily realize that with so much of Infinera’s valuation riding on its performance with metro products, they will be even more inclined to strive to get even better pricing deals, obviously resulting in lower margins. In addition, instead of Transmode’s products originally operating in its traditional environment of relatively limited pricing competition, consistently resulting in margins over 50 percent, they will now be sparing in a more aggressive arena. In fact, as we indicated in an earlier post, at last year’s webcast, Insight Infinera 2015, there was favoritism being shown to the internally developed metro solutions over those from Transmode, resulting in an inevitable amount of internal cross-competition, and one could surmise that it is a possible explanation as to why a major customer is holding up purchases. More critically, on the earnings call, with the statement, “there is an opportunity that’s been driven to them by one of their customers and I think they get back on track soon,” could suggest an OEM arrangement, which could certainly mean the possibility of additional losses moving forward.

Regardless of the type of buyer in that particular situation, all network operators that have Transmode equipment in their infrastructure have to be concerned about the level of support and development of these products in the future. They also are probably alarmed about all of the discussion surrounding changing the fundamental nature of these systems with the incorporation of PICs. The most symbolic action of the squandering of the opportunity with Transmode was that despite the past assertion that Infinera would not disaggregate the level of performance on the metro business, when there was a decline in sales, the Americans did not hesitate to throw the Swedes under the bus.

As we pointed out in a recent article, it is hard to believe that Transmode engineers could have failed to have developed a product by now that would have better competed, especially in terms of density (which is constrained with the PIC approach), with the new solutions being introduced to take market share from the CX. Perhaps the handwriting was on the wall when Infinera felt compelled to put out a release on DCI business involving a tier 2 service provider, Windstream, while the bulk of the action is of course in the hyperscale realm.

Speaking of at least some of the hyperscale companies, we believe that Infinera may have been disingenuous in the past about their clear intentions to migrate towards white boxes or open line systems, which is certainly not been in the best interests of the manufacturer. Concerning, the contention in a September 28, 2015 Investor’s Business Daily that the “gross profit margins in the optical industry are in the low 40s, compared to 70% in the networking switching market [and thus] optical gear is less vulnerable to ‘white box’-style competition,” the vendor’s historic association with Google and its manic cost reduction efforts across the food chain should have indicated to the supplier otherwise. When Infinera claimed: “It is very simplistic to view the [data center] market that I’m going to go trust my life on a white box,” it was ignoring the fact that these systems do not stay in these networks that long, as well as their protection mechanisms.

Returning to ICE, which Infinera states will have an impact on all of its product lines, as we talked about in the past, given the high-level of investment in a new PIC, it has to go for the home run, if you will, the highest amount of total bandwidth at the longest reach possible. Given another matter that we have also addressed previously, we are not surprised that with the expected limitations of an Indium Phosphide chip, it did not go beyond 16QAM. Therefore, it is constrained by having to go with multiples of lower data rates, which we considered a potentially great challenge to pull off at a terabit, let alone 2.4 terabits, and helps explain why it put off the timing of availability question until this year’s Analyst Day.

On a closing note, Ciena’s high-level engineers never lost any sleep over the years about Infinera because the former always considered the latter as intrinsically as an LH play, regardless of any previous rhetoric. Taking everything into account, this characterization could easily remain the case, and if so, the long-term implications for Infinera’s future growth are quite evident.

[written by Mark Lutkowitz]