Thoughts on Juniper Networks

If you read research reports on Juniper (JNPR) and you did not have an extensive background in networking technologies, it must be difficult to figure out all the different perspectives.  Here is how I would think about JNPR, but I am writing this from the notes I took reading through the last ~6 months of reports.

1. Prior Posts: I think most of my past posts on JNPR can be found here.

2. Product Cycles: Long time readers know that that I believe technology markets and cyclical and positive product cycles are fundamental to positive equity performance.  This not only includes revenue growth through market share gains, but also sustaining and expanding gross margins.  The fastest way to kill a tech stock is to shrink gross margins.  With that in mind there are numerous reports and a recent upgrade based on JNPR entering into a positive product cycle for 2012 with (i) T4000, (ii) PTX and (iii) QFabric.  Last month John Chambers said that “Juniper is the most vulnerable I’ve ever seen.”  Which one is it positive product cycle, vulnerable company or somewhere in the middle?

3. PTX: Conceptually PTX is an interesting product and with all the OTN offload, router bypass, hollow core, transport router talk over the past couple of years, I can see why JNPR built the platform.  On the other hand, the market for the additional capabilities is orthogonal to JNPR’s core business gross margin (GM) level.  JNPR and CSCO are companies that enjoy a premium gross margin advantage.  Traditional transport companies (e.g. CIEN, TLAB, ALU, ERIC, ADTN, INFN) regularly report GMs about 2000-2500 bps lower that CSCO and JNPR.  What I do not know, but look forward learning is how the PTX will be priced and purchased in the market.  Can JNPR maintain their traditional GM level in the mid 60s for the PTX platform?  Will service providers grind them down?  That is the conundrum I am curious to learn about.  A company like CIEN would enjoy adding some packet capabilities, but not a full router, to their product and receiving 500-1000 bps more in GMs.  I am not sure JNPR would like selling their PTX for 1000 bps lower than their traditional GM level.  I have no idea which way it will go, but I would say it is easier for the transport companies to add some packet and live in the market at the 40-55 GM level than it would be for the router companies to add some optical transport and live in a market of 50-55 GMs.

4. QFabric: The CSCO team is quick to point out three failings with QFabric: (i) it is proprietary, (ii) failure domains are too large and (iii) complex software.  Of these I am will to believe one and half.  The intended nature of the fabric design and evolution from merchant silicon to an ASIC is to run cells between the ToR and the fabric.  CSCO says this is proprietary, yet they are running TRILL for Fabric Path and it looks like you can only connect their ToR to the Nexus 7k and you a need software upgrade.  What is the difference?  TRILL is quasi standards based and JNPR cells are proprietary?  This is slicing hairs and is really meaningless.  Both vendors are locking out other ToR suppliers from connecting to their fabric architectures.  As for the failure domain being too large, I am not buying this.  As for the software being complex, I think that is entirely possible.  Will this solution be a big seller?   I am not so sure – but it is a new approach.  I think JNPR has some deliverable issues to overcome and the solution becomes more complete in 2012 with version 2.0, therefore I think the assumption of a large revenue ramp in 2012 is overly optimistic.

5.  Trading Range and Chart: In September I told a PM in NYC to buy JNPR. The stock was about $18 and had yet to bottom.  He asked me why, knowing I had been negative on JNPR.  I told him it was just a trading call at this low level as and bunch of tech trading desks would get positive on it and we would get goofy notes from analysts who are positive on JNPR saying the selling had gone too far.  That is exactly what happened.  I updated my JNPR weekly chart that I posted a few months ago.  I circled the high and low.  Before looking at the chart, do you know where JNPR is trading on a chart over 123 months?  The answer pretty much the middle.  The median price point over that time is $24.58 and the stock is $24.92 today.  If you look at the chart, you will see it spends time around the $24.56 level, which is the 50% fibb level from 2002 low and 2011 high.

6. Upgrades and Downgrades: The following chart is just a list of ratings changes I copied off Dow Jones.

With all that being written, it is possible I am mistaken and got it all wrong.  🙂

/wrk

* It is all about the network stupid, because it is all about compute. *

** Comments are always welcome in the comments section or in private.  Just hover over the Gravatar image and click for email. **