A progress report on Infinera for Q1 - part 1


I have been preparing some notes for both review and reflection. The notes below are my attempt at a "progress report” that summarize Infinera’s first quarter. Much of the detail behind the summary was taken from various former posts and explorations.

To make the information easier to read I decided to organize my notes by each of the major product segments: DCI, Metro and Long Haul. I also decided to tackle DCI first, and now that the DCI part is ready I can share those notes with you.

I do have one request though before you continue. I have every intention of completing a summary for both Metro and Long Haul, and hope to do so before the Q1 earnings release. One thing I’ve found, however, is that it is much harder to “summarize” all the info than it is to just report it. Please bear with me!

Here’s DCI:

DCI (Data Center Interconnect)

Key metrics for the product segment:

  • Infinera is expecting double digit growth in their data center interconnect segment in 2016.

  • Their estimated TAM in this segment is approximately $828M for 2016. Their TAM in 2015 was $552M. They estimate a CAGR of 43.7% from 2015 through 2019 - or $2.4B in TAM at the end of the last year in their current forecast.

  • Infinera models their TAM at 50% of what the market research firms in the industry assign (research firms roughly expect a $5B in TAM by 2019). This is because Infinera’s application in DCI (with Cloud Xpress) does not address the entire DCI market, and knowing this, they had modeled their projections accordingly. See page 25 in http://s21.q4cdn.com/892601718/files/doc_presentations/2016/… for more info.

  • This segment of their business is still small, but growth is accelerating rapidly per recent conference call notes and press releases.

Key observations:

  • DCI has only just started its ramp toward the end of 2015, though the product has been shipping from over a year ago. This was because the field trial and certification ramp required by customers turned out to be lengthy one.

  • However, the long window for certification can also bring certain benefits. Other vendors who wish to sell competing products will experience the same lengthy trial. As a result, Infinera currently sees virtually no competitors selling product in their TAM in any material way. If that trend continues they will consume a very significant portion of the $828M TAM calculated for this year.

  • Products currently in trial will also be measured against Infinera’s results. Beating the high expectations established be the incumbent won’t be trivial. It was disclosed during a recent Infinera analyst/investor’s conference that Ciena’s DCI trial wasn’t going so well. This may be why.

  • A high number of customers for Cloud Xpress also wanted to wait for Infinera’s 100Gbe version last year. This also pushed out the timing of the ramp, but this also suggests their customer base is either building very advanced networks, or they are planning for the future, or both. This will be another barrier for a new entrant to breach if they wish to sell into Infinera’s market.

Expectations for Q1:

As far as what to expect from Cloud Xpress in Q1, we first we got an update on that right in the middle of the quarter during the Q4 conference call.

“The momentum that was building near the end of the year has actually accelerated in Q1 for the Cloud Xpress product. It’s actually in Q1 so far exceeding our expectations, and that’s what we thought would happen, it’s architectural win, it’s a proof point in the industry and even though a number of competitors have announced competing products, we are really not seeing them impact the market yet.”

We also got a few PRs on Cloud Xpress deployments through the quarter, covering EMEA, North America and APAC:

GEANT: “The GÉANT network is one of the most advanced research networks in the world,” said Nick Walden, senior vice president, EMEA Sales at Infinera. “Cloud Xpress allows GÉANT to continue to leverage Infinera’s Intelligent Transport Networks to power its network – from the metro NREN connectivity to the backbone.”

Windstream: “With our initial deployment of Cloud Xpress, we are able to meet the growing bandwidth needs of our customers needing connectivity to high demand Internet Exchange facilities like 350 Cermak,” said Joe Scattareggia, Senior Vice President of Sales for Windstream Carrier Solutions. “Cloud Xpress delivers the space savings, low power and high bandwidth required to serve our wholesale and enterprise customers with optical networking solutions at bandwidths up to 100 Gb/s.”

NHN Techorus: “Techorus’ selection of the Cloud Xpress further reinforces the value of Infinera solutions for data center interconnection across the globe,” said Andrew Bond-Webster, vice president, regional sales APAC at Infinera. “Infinera continues to see traction among our customers deploying Cloud Xpress to scale capacity for data centers. We are especially pleased with the success we have seen providing Cloud Xpress to Techorus and other customers in Japan and the Asia Pacific region in conjunction with our highly valued partner, Nissho Electronics.”

Last, we received some future looking insight from David Welch during an interview with Light Reading last week. He spoke with regards to DCI growth going forward into the year and the next. He was asked specifically about potential impact from short reach direct-connect solutions to their DCI business. His response: "We’ve got our projections for our business and this doesn’t change those projections at all.”. Looking at how they deliberately calculated their TAM model I can say I believe him.

Some additional points of reflection:

At the end of last quarter it was noted that Infinera had been building excess inventory. Their reason for doing this was to be in a better position to address their customer’s needs and shorten the deployment cycle. They noted on the conference call that some of their customers are starting to buy on win, meaning, when they win a large deal then they go out and buy the gear. The excess inventory build was designed to address that new type of behavior.

Google has been revealed as an Infinera customer by one of the analysts. The analyst also cited Cloud Xpress as Google’s preferred DCI metro product. Google has recently announced some very large and rather significant wins on their platform that includes Disney, Apple, Home Depot and Macy’s (https://medium.com/@jrodthoughts/how-can-google-cloud-win-in…).

It was also noted in the Q3 2015 conference call that one of their customers had decided to push their Q4 order into Q1. In the Q4 call more information on this order was revealed. Tom said it wasn’t necessarily a large order, but he made the comment because it was by someone prominent in the industry, someone you pay attention to and when they do something different you take notice. He made the remark about this becoming a new pattern to watch out for. If the order being pushed out into Q1 was the genesis for the new "buy on win” behavior it could be the reason for building out the excess supply.

Concluding remarks:

From all that I have read and gathered I feel confident about Infinera’s strength and position in DCI and I believe they have a fairly significant first-mover advantage. I look forward to seeing how this matches up with their conference call update coming out next week.



Thank you as always, Kevin. Your careful and diligent updates help me greatly.

…building excess inventory. Their reason for doing this was to be in a better position to address their customer’s needs and shorten the deployment cycle. They noted on the conference call that some of their customers are starting to buy on win, meaning, when they win a large deal then they go out and buy the gear. The excess inventory build was designed to address that new type of behavior.

Having been, in another life, a tech buyer and manager, I think this is a canny move on INFN’s part. I was buying commodity items like Thinkpads, so my issue was which vendor to deal with. I quickly ‘fired’ all the vendors that couldn’t ship product promptly; I could always get what I needed from another vendor, and I developed a short list of vendors who could deliver. I also changed the platforms we supported in favor of another manufacturer that could provide better support for their product.


There’s little more aggravating than evaluating the prospective products, running the trials, making the buy decision and pulling the trigger, only to have… nothing happen. For too long.

IMHO, this “ready-to-deploy” inventory strategy will turn out to be another win for INFN. Who’s going to wait for a solution that’s “good enough” when they can get a better product sooner?