Our initial thoughts on Insight Infinera 2016, as outlined on our LinkedIn fibeReality page over a week ago, was partially that “time will unlikely be a friend [to the vendor] as questions over long-term survivability are apt to become more pronounced.” We also stated that “[t]he only edge that we have been aware of over the years has been the ease of turn-up.” Actually, in having all of the DWDM channels built in on a card, such an advantage would be the bare minimum expected. At least the vendor made an attempt to clear up one of our questions in our last blog post on Infinera showing trans-oceanic growth (when it is not a full turnkey vendor), by stating it is about trends across the board.
Overall, we found the long presentation to be at least somewhat delusional with extensive development continuing on its PIC approach (that might even require moving on to a different technology than Indium Phosphide), which will only get more complex in getting up to higher capacities, despite the supplier finding itself in a situation when it does not see a lot of short-term opportunities for growth. Moreover, Infinera is developing new products with the frenzied hope that something will stick. Ten of the most puzzling outcomes of the webcast are as follows:
1. Overly exposed to long-haul market: Hardly a revelation in that Infinera has talked about making further penetration into the metro space since its inception. Of course, the botching of the Transmode acquisition did not help.
2. Greater amounts of liquid bandwidth/sliceability: Even when it came to its 500G capability, there were not totally clean breakdowns in terms of the price per additional 100G, and this situation could become even more acute in licensing an even greater number of 100Gs. In addition, as we addressed in our post on Xtera, there can be a reluctance on the part of service providers to go for such a scheme in that they are always wondering if they are at least partially paying for extra capacity that they are not using at present.
3. Courage to make key investments: Not really when it is using money from shareholders. Infinera is committing to widespread deployment of 2.4 terabits and development of 9.6 terabits in the short term in which R&D will reach the mid-to-high 20s percent of revenues. How is that level of spending sustainable and not an act of desperation? Continuing on with a technology approach for so many years that has not resulted in a compelling differentiator is not bravery, it is called something else.
4. No idea about next quarter: To back up Infinera’s recent assertions, the executive from Telia Carrier said that it was impossible to tell the added capacity needed in networks in the following quarter. We do not believe that passes the laugh test (certainly, his direct superiors at the operator would not accept such an answer). While Infinera finds it “amazing” that 10G services continue to grow in future years (despite the large continuing business for 1Gs), it appears to have a relatively high level of misplaced confidence that the transition to 100G exceeding 10G will happen in the foreseeable future, despite contending that it is totally clueless of capacity demands in a subsequent quarter.
5. Should not expect big M&A: Not exactly a surprising revelation, and so mentioning it was superfluous. Infinera has never been a major acquirer of assets outside the company. With no other vendor following its technology model, what would be the point?
6. Looking to be in the top-three DWDM vendors worldwide: Overtaking Ciena, Nokia Alcatel-Lucent, and Huawei does not seem logical. Depending on the cyclical nature of the business, being ahead of ZTE or even Cisco Systems at a particular time may be a stretch.
7. Technology gap in subsea: Curiously, Infinera did not elaborate on this point. However, with doubts over its long-term viability, we fully expect Ciena to continue to dominate the undersea upgrade market.
8. Metro much slower process than assumed: Either the Infinera management is being disingenuous or is admitting to being inexcusably ignorant. Again, the supplier is not exactly inexperienced with the dynamics in the metro space. We wrote an article a while back that Infinera should lower the expectation level regarding this sector.
9. Meshponder for long haul and subsea: Really meshes are unique to the metro space. Putting that aside, Infinera is suggesting that there will be great needs for improving the efficiency of even 200G and 400G routes. Even if we were to assume a critical mass of circuits beyond 100G, which we do not in any kind of foreseeable future, why did it appear that the issue of dealing with stranded capacity was not directly mentioned, which is a big problem with just 100G?
10. Distorting rhetoric: The 500G capability “did not [shake] the industry.” Despite the need for 2.4T being what we called “a last-ditch” effort, Infinera is doubling down, and referring to it as coming out about a year later than it should have been – an interesting reverse psychology kind of trick. The notion of ten out of the top Fortune 100 companies are involved in heavy investment in cloud services does not negate that less than ten percent of the data centers are interfacing at 100G rates. Lastly, the idea that its technology “protects yourself against higher QAM rates” is just another suggestion that there are limitations with Infinera’s devices, and while the industry going in a big way beyond 16QAM is probably a stretch anytime soon, we would not be surprised if the supplier would automatically be eliminated from consideration by engineers who want to cover themselves in checking off that box of perceived requirements.
Please consider our latest reports, Clash of Metro 100G Optical Vendors with Shifting Network Paradigm and Clash of Optical Component Vendors & Technologies in Data Center Networks. For more frequent updates, please follow us here.
[written by Mark Lutkowitz]