It's always fascinating that companies can totally change their operations/business models over a very short period of time. SCMR has been written up on VIC in the past but the thesis now is very different from the last write-up. Last time the stock was trading below net cash. Today, the company trades slightly above net cash ($16/share in cash) but with a cheap option on a potentially revolutionary new software product. Plus, we have a management team that has very effectively allocated capital in the past--for years, they have recognized that optical products are totally commoditized and any investment in them is a poor use of capital as we can see by CIEN or JDSU's lack of free cash flow generation and very poor returns on capital. They have alluded to returning all of the cash to shareholders if they can't get any traction on their IQStream software product, which seems very unlikely from our checks. 

 

To sum it up, IQStream (IQS) is a new product in trial testing with multiple wireless carriers which reduces congestion in their wireless backhaul networks. This is huge problem for wireless carriers that has resulted in capping data usage...etc. Based on the assumptions below, if IQS works it can conservatively deliver $1-2 in EPS next year on just one or two carriers, but this is only the tip of the iceberg. There's a lot of buzz around the product and if it works, it will likely be adopted by many carriers across the world which would provide massive upside (false precision to try and quantify, but it would be big), $5-10 in EPS potentially. On the downside, if the product does not work, we will get an untaxed $15-16 back, minimizing downside from the current stock price. In a highly correlated market, this seems like a pretty good risk/reward to me.

 

The way I think about near term EPS based on the info below is if VOD Europe alone has 50k base stations at which IQS can be implemented and the company tries to get a 50% GM (along the lines of their corporate average), they will sell the product at about $6k/base station. Assuming they choose only to implement in 10k base stations for the sake of being conservative, that’s still $60m in revenue and $30m in gross profit. Since opex is minimal and they aren’t paying taxes, let’s assume 35% of revenue flows to FCF, that’s about $1 in cash EPS for 1/5 of VOD’s European base stations. They are also in trials with 4 other operators. Even if only 2 come through, that could be $2 in cash EPS. On a $19 stock with $16 in cash and $3.50 in NOLs (though we don’t want to double count since we are assuming 35% cash flow margins), that $2, even at a 10x multiple would mean an additional $20 to the stock. So, worst case, if things don’t work out, they have alluded to closing down the company in the next year. If that happens, we get $16 in an extraordinary dividend (untaxed). Our downside is $3 but our upside could be $20+. We have downside protection here and the ability to participate in massive upside should this solution work out.

• Sycamore Background

-IQStream (IQS) is a departure from what they’ve previously done

-Innovator in smart optical networking in 98-2000, had a unique offering but the business became commoditized very quickly as larger operators at scale developed comparable products and cross sold through larger suites to carriers

-SCMR still maintained carrier relationships and continued to sell commoditized products, putting only enough R&D in to stay competitive

-Company has been dormant for several years but checks have shown Vodafone was looking to optimize their network solution (had an RFI with 10 vendors ranging from offloading video optimization to streaming/shaping technologies) when SCMR ‘knocked on their door’ with a very different solution, unlike the others, a data deduplication system that’s similar to WAN optimization like Riverbed or Bluecoat (these cos help enterprise customers accelerate access from branches w/in private networks)

 

IQStream (IQS)

Sycamore’s technology can compress the traffic and apply it to the wireless backhaul space

-This is challenging bc (1) it’s not a typical IP based environment and (2) backhaul is encrypted between mobile devices and RNC (Radio Network Controllers) at the head end of the backhaul

-IQS operates in the backhaul space and has implemented large scale encryption capabilities and maintains the user session while maintaining the handover

-Encryption and bookend caching are the major barriers to entry here

 

Caching solutions are at the heart of the product

-Others have looked at doing a simple cache like LLNW or AKAM, but conventional caching just sits at the end of the network looking for requests, i.e. sees HTML for a website already in the cache which saves network bandwidth

 

This single cache system is hard for wireless backhaul

-In wireless backhaul it is hard to inject cache into the RNC and the Node B (Node B in basestation) since the RNC sits further back in the core of the network and controls 150 base stations—hard to dual cache from one RNC to many Node B’s

-Very tight relationship between the Node B and the RNC, so it’s hard to cache/sit in between

-This problem doesn’t really exist in 4G networks because there is no RNC so LTE networks are not as tightly controlled

-SCMR is addressing the problem of today—so this is not a solution with revenue a few years down the road, the addressable revenue opportunity is today so revenue should be coming in very soon

-Bottom line is that the technology allows something like caching but not quite normal caching

-In IQS, there’s a large device co-located within the RNC which works with a number of smaller devices in each Node B

-Each node supports communication protocols, it takes a pair of devices and spooks the RNC into thinking it’s talking to the Node B directly. It looks invisible, just a bump in the wire and is not visible to either device so it just passes tokens over the backhaul network

-Backhaul is the largest capex/opex expense for the operators

-HSPA has stepped up throughputs but unit costs in backhaul bandwidth are not coming down

-The alternative is to push fiber to the backhaul but that’s very expensive except in urban areas—as such, there’s a strong desire to find innovative ways to upgrade

-There have been improvements in microwave radios to increase bandwidth of microwave networks but not enough to make a big difference

 

Back to IQS—here’s how it works

-Everyone in a certain area watches a viral video on youtube from their tablets or smart phones and it becomes a heavily watched video

-W/out IQS this results in individual streams to users from their wireless carrier which is very bandwidth intensive

-IQS recognizes a bit patent it has seen before the second time the video is viewed. The RNC box recognizes this. The base station box also has this stream in its cache and it recognizes the bytes since it has it in its cache. As such, it doesn’t need to re-send the data over the backhaul, just a small token to remind which bit patent was sent across the backhaul preivously for the video

-This is a book ending approach which spoofs the RNC and the Node B into believing they are talking to one another which a normal caching system can’t do—otherwise the RNC and Node B would get out of step since a conventional cache can’t pair devices

   

Trials with Carriers

-SCMR has been very successful so far and the benefit here is that through the trials we’ve seen that interoperability works with IQS because of spoofing the RNC and Node B into believing they are talking to one another

 

Revenue/TAM

-Despite this being used mostly for 2G/3G, it will likely be used in base stations for 10-12 years given the amount of time required to build out LTE

-There is a role for IQS in 4G too but it will be easier for competitors to enter the space for the reasons listed above

   Movik Networks—VC startup trying to do something similar for LTE

-Cell sites are classified into categories as to where IQS is feasible to use

• No need in urban areas where backhaul fiber exists, no need in rural areas where there are no bottlenecks to bandwidth
• This will be used mainly in suburban areas and the outskirts of urban areas
• In Vodafone’s German trial, there were 22k base stations, taking out 25% urban and 30% rural leaves a market of 45% of the basestations
• For Vodafone’s 120k European base stations, 45% would mean it would work in 50k base stations

-Pricing: We don't have much color on pricing but we know the Kasumi Encryptions are the most expensive cost at about $500 or so and are needed on both ends. The cost will generally come to $3k/base station—pricing by B Node, thus a 50% gross margin would price it at $6k

-Sycamore is generally not scaled to do worldwide fulfillment, which is why they are currently speaking to RAN vendors and have announced partnerships with them—this will also help them to create a more general product that they can sell directly through each operator

-The costs are minimal to SCMR. Trials are done in 2 steps, lab based trials at minimal costs and field trials—there are not many units out in field trials. Customers have been very cooperative—there were just a few SCMR employees in Portugal when VOD was doing their trials there. The trials were mostly conducted by VOD employees

-We have no sense as to when the trials with VOD or anyone else might finish but given that it's been going on for over a year, it seems like a conclusion, one way or another, is imminent in the next few months.

-From our checks we are seeing some new technologies to help with challenges of congestions but none of them are competitive with IQS for backhaul in 3G networks

-Congestions happening today is already a problem and carriers are proactively pushing for a new tool in the toolbox and IQS is just that

-Because it’s cheap and can be implemented per base station, the value here is that it can tactically be pushed out over a small number of base stations to start so SCMR can begin to see revenue from many operators for a small number of base stations

Annoucement of IQS adoption by Vodafone or one of SCMR's other trial partners.