Description

Ikanos Communications, Inc.

“I think it’s a business in an industry that’s really changing. This has historically been about the worst industry you can think of in technology…you had literally dozens of competitors… and what’s happened over the years is it’s become such an awful industry that most people have left it. So then it’s now consolidated and we’re down to three players.”

                                                                -David Einhorn on his recent Micron Technology pitch

This isn’t meant to compare MU to IKAN but the 200% YTD rise in MU and Einhorn’s previous success with STX under a similar consolidation thesis shows what consolidation can do for the remaining players in a fiercely competitive tech industry and how rationale behavior amongst 3 survivors can produce extraordinary returns for shareholders.  However, unlike MU and STX which are both near all-time highs, investors have an opportunity with IKAN (down ~24% YTD) to get in before the benefits of the consolidation in its industry and more importantly right before the beginning of a significant multi-year growth period in its addressable market is realized.

Executive Summary

Ikanos is an unprofitable fabless semiconductor company specializing in DSL processors that power telecom carrier infrastructure and customer premise equipment.  It has experienced sharply declining revenues over the past couple years due to the secular decline in DSL equipment spending as alternative broadband technologies (cable and fiber) take market share.   The DSL industry is however on the brink of a new era of rapid growth as operators around the world begin to deploy a new noise reduction technology known as vectoring to maximize the untapped potential of their copper infrastructures. 

In an ideal world, operators would obviously prefer to deploy pure fiber networks or fiber to the premises (FTTP) but the reality is that capital and time constraints (due to competition and consumer demand) are forcing most to seek out alternative ways to deliver increased speeds.  This dynamic has resulted in more and more telcos choosing to forgo FTTP in favor of FTTx (FTTC: Fiber to the cabinet or FTTN: Fiber to the node) deployments where fiber is combined with existing copper - usually somewhere in the last mile where the time and costs of laying fiber become uneconomical. 

Ikanos was the historical market leader in next generation VDSL technology but has only recently perfected the necessary products to allow telcos to achieve the long promised speeds of 50-100Mbps.  They operate in an oligopolistic market place with 50-60% gross margins and only 2 other players: Broadcom and Lantiq - both of whom having shared in the industry woes over the past few years are also gearing up along with Ikanos for a new "golden age for copper."

And even though Ikanos is grossly subscale relative to its 2 competitors, it has developed an industry recognized and differentiated technology that, combined with its strong and long standing customer relationships, should allow it to win more than its fair share in an industry niche that is expected to grow at a 25-30%+ CAGR over at least the next 4 years.

Due to its disappointing history in an extremely challenged industry, Ikanos has been long forgotten and is covered by only one analyst at Needham who had a HOLD rating until just recently.  I hope to show in the following sections that Ikanos is at an inflection point and offers substantial upside with little to no downside at current levels.  I expect 2014 to be a breakout year for Ikanos and believe there are multiple catalysts that will force the market to re-evaluate this hidden growth story.

History and Why Opportunity Exists

IKAN IPO'ed at $12 in late 2005 and raised nearly $70M.  By early 2006 it had shot up to over $20 and early VC backers and other insiders cashed out over $64M through a secondary priced at $20.75 with the remaining proceeds of nearly $48M going to Ikanos.  The secondary was timed perfectly for insiders as Ikanos almost immediately began its descent, eventually losing over 95% by the end of 2008.  The initial hype around Ikanos was its success in supplying chipsets for the build out of the Japanese and Korean DSL markets and the anticipated future rollouts of VDSL, where Ikanos was presumed to have superior technology or at least a head start relative to its competition. 

The combination of extreme customer concentration (over 97% of their sales going to customers in Asia and over 40% of revenues to a single OEM customer), too many players in an already intensely competitive space, and the industry's total dependence on the fickle deployment plans of telecom carriers was a recipe for disaster.  Sure enough, VDSL deployments were delayed, competitors encroached, and Ikanos never achieved the dominance or even return on its significant R&D outlays that its investors had hoped for.  As the industry consolidated, Ikanos was able to acquire several of its former competitors - all the while burning cash on its R&D efforts to stay ahead of the curve in next gen VDSL technology in hopes that carriers would eventually seek to leverage the potential of their massive copper infrastructures.

In April 2009, Ikanos announced its acquisition of Conexant's Broadband Access Product line for $54M which would be funded by an investment of $42M from Dado Banato's Tallwood Venture Capital.  The deal gave Tallwood 24m shares ($1.75/share), 7.8M warrants struck at $1.75, and two board seats for Dado Banato and George Pavlov which effectively gave Tallwood control of Ikanos. 

Ikanos proceeded to rise to over $3 before falling back to earth and having to raise ~$11M at $1.05/share in November 2010.  Since then, Ikanos briefly touched over $2 earlier this year but was again forced to issue equity at just $1/share raising ~$23M in early November 2013 to make sure they would be in compliance with their term loan agreement.

IKAN's leadership has also been in constant flux with 7 CEOs since its IPO in 2005.  Today, what we are left with is a company that, despite being public for over 8 years, has the unsustainable cost structure of a development stage company with well over 50% of TTM revenue going toward R&D.  They operate in an industry that is notoriously opaque with carriers who keep contract awards secret until final deployment and vendor's who are similarly tight lipped with any design wins to chipmakers.  The chipmakers also have virtually no control over carrier deployment schedules and have historically reported surprise revenue misses as a result of frequent carrier delays. 

In addition, quotes like the following from Verizon's CEO seem to suggest the imminent death and abandonment by telcos of copper networks:

"Every place we have FiOS, we are going to kill the copper.  We are going to just take it out of service. Areas that are more rural and more sparsely populated, we have got LTE built that will handle all of those services and so we are going to cut the copper off there."

A cursory glance at the data would seem to validate this as well: Jeff Heynen of Infonetics noted earlier this year that worldwide spending on DSL equipment plunged 26% in 2012 with much of this due to China's shift from ADSL to FTTH.  This perception that DSL is a dying technology however could not be further from the truth.

Market opportunity

"There was a feeling that DSL was almost done. What happened during the last couple of years was that the use of vectoring -- a crosstalk-canceling technology that was thought to be a very exotic -- became more feasible. Ikanos was one of the pioneers that showed that it can be done for a reasonable cost. During the last year most of the carriers have seen that vectoring can be used for the mass market."                                        - Arun Hiremath of Ikanos in May 2013

Ikanos and the industry have been talking about the theoretically achievable 100Mbps+ speeds of DSL since its IPO but the difference is that next gen technologies like bonding and vectoring have finally been perfected and carriers, after years of testing and trials, are finally convinced and beginning to deploy.  Mr. Hiremath continues:

"VDSL came along in 2004 or 2005 and held the promise 100 megabits per second down and 50 megs up to a half a kilometer range. But when carriers put it in the field, crosstalk drove the performance down drastically. It wasn’t meeting its promise. Vectoring VDSL cancels that crosstalk and allows you to get the performance very close to what you get in a single line performance with nothing else on the line. It’s called effects-free."

NodeScale Vectoring is IKAN’s proprietary solution for achieving this performance and will be discussed in more detail in the competition section.

Lee Ratliff of market research firm IHS reports that:

"VDSL is enjoying a new lease on life after having been nearly written off, with its speedy transmission rates now proving to be an attractive option for telcos compared to expensive fiber technologies...VDSL is set this year to enjoy impressive 40 percent subscriber growth...Robust growth exceeding 25 percent is in store for the VDSL market each year until 2017...Overall, the five-year increase from 2012 amounts to a notable 256 percent."

So how does one reconcile this with the sharp declines in overall DSL equipment spending?  The key is to realize that global DSL subs dwarf all other technologies and then differentiate between the secular declines in older ADSL tech and the growth of next gen VDSL2 and FTTx deployments.  

Source: Point Topic Q4 2012

In the previous section, I omitted that Jeff Heynen also noted that "despite the difficult road for DSL, VDSL remains a real bright spot, expanding among operators in Western Europe, North America and Latin America. Vectoring solutions and a long-term path to G.Fast are driving sustained interest in VDSL2."  He also noted VDSL CPE shipments jumped 39% in 2012 and even though total DSL equipment revenue fell to its lowest level in 2Q13, VDSL maintained its robust growth at 30%+ YoY. 

And with respect to China's well publicized transition from ADSL to FTTH it's been recently reported that even China Telecom, which continues to be the biggest investor in fiber in China, has expressed interest in VDSL2 and in early 2013 completed China's first field test of vectoring technology with hopes of utilizing its existing copper infrastructure to help it achieve China's aggressive national initiative to connect 250M+ urban and rural homes to high-speed broadband by 2015.

With all the buzz around fiber, most would be surprised to learn that VDSL subscriber growth is actually forecasted to grow faster and from a larger base than FTTH over the next 4 years.  IHS and Infonetics market data on actual and projected unit shipment volume is inconsistent but both project CAGRs of 27-30%+ over the next 4 years eclipsing the <20-22% projected growth in FTTH subs.  Management claimed even higher rates on their Q3 Call: "transition to G.fast is making DSL, for the last mile, a very viable and attractive alternative to costly FTTH deployments, making hybrid fiber copper networks and FTTx solutions the fastest-growing segment of the broadband access market, as highlighted in a recent report from Point Topic, outpacing FTTH by more than 2x."

In North America, telcos continue to be challenged by cable operators that are - on average - offering significantly faster speeds and with the introduction of DOCSIS 3.0 the gap has only grown wider.  Ratliff explains, "The consensus is that VDSL has finally found its place in the broadband market, especially in mature territories such as the United States and Europe. And while not a substitute for FTTH, VDSL could fill a short-term niche for telcos scouting for an inexpensive solution as a bridge to a more future-proof FTTH network."  According to the most recent data from the FCC published for 2011 - 69% of cable subscribers are receiving speeds above 10Mbps compared to only 19% of Telco (DSL + FTTH). It gets worse when you remove FTTH and realize only 6% of DSL subs get above 10Mbps.  If telcos aren't able to implement a rapid and scalable solution (both of which are not characteristic of fiber deployments) then they risk not only losing broadband internet subs but voice and video as well.

FTTH, while appropriate for certain markets, is for the most part overkill for today's consumer demand.  It is incredibly costly (anywhere from 5-10x more expensive than VDSL) and takes years to fully build out.  In contrast, VDSL Vectoring, can be deployed in a matter of months, and with speeds of  50-100Mbps will be capable of satisfying bandwidth needs for the next 5-10 years.  Unlike China and other countries that offer government subsidies for laying fiber, the rest of the world has seen fiber deployments abandoned due to lengthy and uncertain ROI models (more on this later).  Ratliff remarks that "the turnaround for VDSL has been dramatic, as the technology was considered a disappointment five years ago and had appeared headed for a shut-out by then more competitive FTTH.  However, the global economic crisis shortened ROI horizons for carriers, delaying FTTH deployments and giving VDSL vendors precious time to perfect vectoring technology."

Verizon's FiOS roll out and early wind down is a perfect example of the challenges with justifying FTTH deployments in the US.  In contrast, AT&T re-emphasized its commitment to VDSL2 with the announcement of its Project VIP in November 2012 and plans to extend U-Verse FTTN coverage by 8.5M homes by 2015.  Another massive fiber deployment disappointment has been the roll out of the National Broadband Network in Australia.  In Europe, Belgacom, KPN, British Telecom, France Telecom, Swisscom, Deutsche Telekom, Turk Telecom, Telekom Austria, and others have chosen to forgo pure FTTH in favor of deploying FTTN using VDSL2.  Deutsche Telekom plans to increase VDSL coverage to 65% by 2016 and Telecom Italia will deploy fiber to the cabinet to 6.1 million homes in 100 cities by the end of 2014. 

Lee Ratliff concludes that "VDSL with vectoring answers the threat posed by DOCSIS 3.0 on the one hand, while also avoiding the expense of a costly deployment previously associated with FTTH" and that for VDSL "the stars have finally aligned—and the technology appears set to stay for some time to come."

With vectoring, the US and the EU finally have a viable and economical way to achieve their bold national broadband plans - the FCC has set a goal of 100M homes with access to 100Mbps by 2020 and the EU's Digital Agenda has targeted 30Mbps+ for all EU citizens and 100Mbps+ for 50% of households by 2020.  The EU is only 2% of the way to achieving their 100Mbps goal according to their Digital Agenda Scoreboard which explains the mad rush to deploy ASAP.  Bill McDonald, director of carrier development and relations at Ikanos, has confirmed that the general sense is that telcos will begin deploying on a massive scale in 2014 with some already deploying in late 2013.  BroadbandTrends also expects, based on numerous service provider interviews, that the bulk of upgrade activity associated with VDSL2 vectoring will begin in 2014.

Critics of VDSL2 say that FTTN deployments are only delaying the inevitable fiber future.  They contend that it’s better to spend the money now since FTTN is only a band-aid.  These people clearly have no appreciation or understanding of the ROIC and time to market implications of reaching their fantasy land idea of the future.  They look at sexy Google Fiber projects and wonder why everyone doesn't just follow their example.  ADTRAN, an equipment supplier for both fiber and copper infrastructures, said it best: "The all-fiber future belongs to those who survive long enough to reach it"

No one disputes that fiber is the ideal but it’s the path to this fiber future that presents an incredible opportunity for Ikanos.  Simply put, the economics for most telecoms of rolling out FTTH for every one of their subscribers simply doesn't make sense right now.  If you're already convinced, you can skip this next part but I think it's important to understand how telecom carrier executives have to think about their future deployment plans.

Deploying a full fiber network is a huge capex endeavor that requires several cost and ROI assumptions to justify - unfortunately for most carriers many of these assumptions are near impossible to account for.  Consider the evolving ways people consume media and the recent trend toward OTT distribution (Netflix, Amazon Instant Video, Aereo, etc) - if cord cutting continues and becomes the norm then carriers can't rely on revenue assumptions from IPTV - which comprise a significant percentage of their ARPU - to justify their fiber build outs.  The carrier would basically be subsidizing the businesses of all other content providers/deliverers by providing the necessary bandwidth to the consumer but not reaping in the rewards of this major capital outlay.  This is why there is such a hot debate going on right now in the media world with everyone fighting for their share of the pie.  These discussions are still evolving and it will be impossible to accurately forecast returns without more clarity here.  Although there is a very small percentage of the population that will pay up for fiber, the mass majority of people simply care about the speed of their service and are agnostic to the medium of delivery.  So if telecom carriers can provide comparable speeds of up to 100Mbps+(which will be more than enough for the mass majority of the population at least until 2020) with VDSL vectoring at a competitive price; then telcos will be able to gradually extend fiber closer and closer to the home while maintaining and even taking share from fiber and cable providers all while earning a greater ROIC for shareholders.  The gradual nature of these fiber extensions is another opportunity for Ikanos and VDSL with FTTdP (Fiber to the Distribution Point) deployments which will be discussed below.

"Over the last year we’ve see the lines blur considerably between copper and fiber. It used to be that our customers were firmly in either a fiber-to-the-home or a VDSL mindset. Now our fiber customers are looking at copper again, and our copper customers are looking at bringing fiber closer to their subscribers – all with one objective in mind: to connect more people, more quickly. And of course the end-user doesn’t care what technology they have. They just want fast internet access, now."

-Dr. Stefaan Vanhastel, Alcatel-Lucent Fixed Networks Marketing Director

FTTdP and G.Fast

There is a new standard currently being finalized known as G.fast that will further extend the life of copper infrastructures with potential for up to 1Gbps through what's known as FTTdP or Fiber to the Distribution Point.  Ikanos actually just received its first OEM win for its Velocity-Uni chipset for use in FTTdP deployments on 11/18/13.  FTTdP will require telcos to install fiber closer to the home than VDSL2 vectoring - as close as 100 meters - in what is being called a hybrid-fiber-copper network. 

Ikanos, Alcatel Lucent, and others have said telcos tell them that deployment costs can be reduced by up to 50% by not having to install fiber over the last 100 meters.  The two major installation problems over last 100m are first outside the residences with digging trenches through lawns and under sidewalks and stringing wires from utility poles and inside MDUs and then inside the residences which requires sending a pricey technician inside the home to run fiber to the modem/gateway.  A major European telco has said that as many as 50% of potential FTTH subscribers had cancelled because they did not want someone running wires into their homes.  Then there's the problem of scheduling a time when the subscriber has an adult present while the work is being done.  Again, G.fast is not a competitor to FTTH but it will allow telcos to further postpone the dates when they have to deploy FTTH.

I see their recent win as proof of their leadership in this next phase of VDSL technology rather than an indication of any imminent revenue contribution.  Large scale deployments probably won't be until at least 2016 but I'm hoping that once vectoring takes off and Ikanos is rerated in 2014 that growth investors buying into the story will apply some rosy assumptions to that opportunity.  Either way, this is not vital to the thesis.  Still someone is clearly interested as OEMs don't make products without a carrier lined up (France's Orange and Telecom Italia are amongst the operators that have looked into commercial deployments).

Competition

Several years of carrier delays in VDSL, long and expensive R&D, and the rapid decline in ADSL forced the DSL chipset industry to consolidate over the past few years and has left only 3 major players: Broadcom, Lantiq, and Ikanos.  In November 2012, Moody's cited similar factors as reasons for the precipitous decline in Lantiq's revenues: a)reduced investment spending by telecom carriers in a muted macroeconomic environment; b) lower-than-expected revenues generated by new product launches; c) lower demand for older-generation products.  These challenges affected the entire industry (with no noticeable effect on Broadcom due to its size and diversity of businesses) and crushed both Lantiq and Ikanos.  Ikanos revenues were cut in half from a peak of over $190M in 2010 to under $94M (TTM 3Q13) with its stock dropping from over $3 in 2010 to current levels.  Lantiq (formerly Infineon's wireline business) was acquired in November 2009 for EUR243M by PE investor Golden Gate Capital - who saw an opportunity to capitalize on the anticipated growth in next gen VDSL equipment - but has since had to inject nearly $200M in additional equity to support the struggling chipmaker and prevent debt covenants breaches. 

However, as covered in the Market Opportunity section, "the stars have finally aligned."  There will of course be headline contract/design wins announced in the coming years which will result in some bigger winners than others in this oligopolistic marketplace but the rapidly growing pie and the dynamic of vendors and carriers not wanting to be beholden to a single dominant chipmaker (like Broadcom) should allow for all three players to prosper and at the very least recoup the significant R&D costs that have been sunk into the development of these technologies.  So while the following analysis on IKAN's competitive advantages may be the difference between a "2 bagger" and a "3 bagger" or more, I don't believe it’s as important as realizing the seismic shift in demand as carriers will be adopting this technology in their deployment plans all around the world. 

IKAN is the not just the only "pure play" in public markets to ride the coming wave of VDSL vectoring deployments but it is also arguably the most technically advanced in vectored VDSL.  While Lantiq has grabbed the early lead in VDSL vectoring by being the fastest to market with a system level vectoring solution and Broadcom offers the most integrated chipset as it has the largest IP portfolio of complementary technologies, Ikanos has superior vectoring technology which is what ultimately will dictate bandwidth speeds.  Broadcom and Lantiq have both touted their system level vectoring capabilities but neither can claim true node scale vectoring which has differentiated Ikanos. 

Competitive Advantage - NodeScale Vectoring

Ikanos was the first chipmaker to introduce and demo vectoring technology coining their innovation NodeScale Vectoring (NSV) in October 2010.  The patents necessary to develop vectoring technology were held by ASSIA, which was founded in 2003 by Dr. John Cioffi, a Stanford professor who is credited with much of the foundational technology used in DSL.  Ikanos entered into patent licenses with ASSIA in September 2009 which was about a year ahead of its competition.  Although NSV was an impressive innovation and proof of concept at the time, it really only gave bragging rights to Ikanos and its vendor partner ZTE who demo'ed it on their DSLAM equipment because service providers were still uncertain about their plans for utilizing copper's potential versus laying fiber.

Still, the bragging rights to "industry first" vectoring innovations continued and as the technology improved so did carriers' willingness to experiment with lab and field trials:

In September 2011, Broadcom announced the "Industry's First Central Office VDSL2 Chipset With Integrated G.Vector Technology"

In January 2012, Lantiq announced the "Industry First VDSL2  Vectoring Chip Enabling Full System-Level Crosstalk Noise Cancellation"

In October 2012, Ikanos announced Velocity™-3, "the Industry’s First Fully Integrated Vectored VDSL chipset Solution"

Contrary to the above timeline, Broadcom is actually the furthest behind with respect to their VDSL technology as their "industry first" vectoring solution in 2011 was only capable of line card or board level vectoring which can only achieve a fraction of the performance gains of system-level vectoring.  Broadcom actually didn't introduce their first system level vectoring chipset until October 2013.  

In September 2011, Ikanos partnered with ASSIA, who has expertise in managing over 58M lines of DSL/VDSL, and together they demonstrated the "First Solution for Accelerating Commercial Deployment of Vectored Networks."  Lantiq followed more than a year later with its own collaboration with ASSIA in October 2012.  Lantiq definitely deserves kudos for beating Ikanos to market considering they lagged Ikanos in both the initial development of vectoring and in partnering with the ASSIA, the industry's premier solutions vendor and network manager.  IKAN's longer time to market would seem to support the notion of additional complexity and superiority of IKAN's NodeScale Vectoring solution but it could just be better execution on Lantiq's part.

IKAN's Velocity-3 chipset with NSV recently beat out its competitors' solutions and won the "Best Broadband Enabler" award at the 2013 Broadband World Forum.  The award recognized Velocity-3 "for its "technical excellence and for its ability to help carriers and service providers increase revenue, manage costs, improve quality of broadband lines, and increase customer satisfaction."  The panel of judges included "industry experts from Deutsche Telekom, Telstra, British Telecom, and Korea Telecom, among others."

To understand IKAN's superiority to the competition's system level vectoring solutions one can start with the white papers put out by different industry participants.  Lantiq's vectoring white paper covers the benefits of vectoring within a single DSLAM or system level vectoring  while IKAN's paper discusses  the more scalable and complex notion of vectoring across an entire node or across different DSLAMs.  ASSIA also put out a White Paper titled Methods for Supporting Vectoring when Multiple Service Providers Share the Cabinet Area where it describes System Level Vectoring (SLV or vectoring across line cards) as only the "first step towards a truly scalable solution for avoiding alien crosstalk" (crosstalk from non-vectored lines or from lines belonging to other vectored groups within the same cable binder).  Per ASSIA, the next step is vectoring across multiple DSLAMs or cross-DSLAM level vectoring (xDLV) in other words Node-Level Vectoring - which is "when the vectored group size is increased to include lines terminating on different DSLAMs." 

Cross-DSLAM level vectoring is a critical factor in maintaining the gains from vectoring when sub-loop unbundling (SLU) is permitted by the regulator (examples below).  SLU is a process by which alternative operators are allowed access to an incumbents copper infrastructure as a means to foster competition and prevent carrier monopolies by allowing multiple operators to compete for subscribers connected to the same local loops.  SLU obligations were adopted into the original EU regulatory framework at a time when vectoring technology didn't exist so now a key issue in regulatory proceedings is whether or not SLU obligations are compatible with the apparent technology requirements of vectoring. 

According to T-REGS, a specialist in EU telecom and internet policy and regulation, formal regulator proceedings and public consultations have been organized by NRAs in Belgium, the Netherlands, Denmark, Ireland, and Italy with informal exchanges occurring elsewhere as of October 2012.  Since then, additional proceedings have reinforced the idea of the feasibility of competing operators deploying vectoring on copper sub-loops.  This is all to say that IKAN's Node Scale Vectoring is the superior solution of the three chipmakers that allows for carriers to achieve the vectoring gains they want and still satisfy regulatory demands for a competitive environment. 

An example of this technological superiority playing out in real world deployments is in the case of Telecom Italia and Fastweb's official memorandum of understanding to pass several million homes with fiber and DSL, often in the same node.  Huawei, which is the #2 DSLAM vendor after Alcatel, is assumed to have the contract and will likely choose Ikanos chips for their NSV capabilities.  Another case was with Deutsche Telekom and their initial claim (that was shared by their historical equipment vendor Alcatel-Lucent) that it's "impossible" to unbundle vectored VDSL to comply with regulator's requirements for competition at the local loop level and still achieve full cancellation.  On the contrary, IKAN's CEO has said that IKAN's NSV chip can do exactly that: "As long as the two DSLAMs are within about 50 meter, we can communicate between them and cancel noise on all lines."  ALU's inability to support xDLV (cross DSLAM Level Vectoring) or NSV likely contributed to their loss of the contract to Adtran.  The full technical differences and advantages of NSV vs. system level vectoring and other work around methods like bitstream access are beyond the scope of this write up and definitely beyond the scope of my expertise but I think these and other data points when taken in the context of historical carrier/vendor (DT dropping ALU for ADTRAN) and vendor/chipmaker (IKAN and Huawei) relationships are significant and point to the advantage Ikanos may have going forward.  These implications are discussed in more detail in the Carrier and Vendor sections below.

Possibly the most convincing data point to substantiate my view that NSV is a real differentiator is the recent open market purchases by IKAN's CTO Debajyoti "Debu" Pal.  Debu has been CTO since 2009 and has never before purchased IKAN stock.  Other than some token purchases by the new CEO shortly after his hiring in 2012, Debu's purchases are the first and only purchases by an executive in all of IKAN's history.  The fact that it’s coming from someone with intimate knowledge of the technical advantages of their flagship product gives me further conviction in my own independent research.  I don't think it’s a coincidence that this is happening right before the expected ramp of multiple carrier vectoring deployments in 2014 and beyond.

Vendors and Carriers

In 2012, BroadbandTrends conducted a survey to determine vendors' "capabilities to support and provide VDSL2 Vectoring solutions" in order to "gain a better understanding of the features supported and experience related to VDSL2 vectoring" and ultimately ranked which were best positioned for the coming wave of vectored VDSL2 deployments.  The results listed Alcatel Lucent, Huawei, and ADTRAN as the top 3. 

Alcatel Lucent was the first vendor to commercially launch VDSL2 vectoring solutions in September 2011.  ALU was able to do this by not using dedicated vectoring chips in its equipment and instead buying programmable chips (FPGAs) which it would then download its noise suppression software on.  This seems to have been a strategic move on ALU's part to try and grab as much market share as quickly as possible and demonstrate itself as the most experienced vendor in vectoring by not waiting for chipsets from Broadcom, Lantiq, or Ikanos.  This is not a long term solution as they're sure to transition to dedicated chips as they will offer greater peak bandwidth, run faster, consume less power, cost less, and have more processing power for larger scale deployments.  Although ALU has successfully established itself as the leader and first mover in vectoring their loss of a major Deutsche Telekom legacy contract to ADTRAN highlights a weakness in their product portfolio. 

Deutsche Telekom had allocated 6 billion EUR for a broadband roll-out in Germany using FTTC + VDSL2 vectoring, making its landline contract the largest contract in Europe by far.  Like other operators, it had slowed its deployment of FTTH, stating that it would deploy only in markets where it is profitable and would require at least 10% of households to commit to the service before they deployed.  Their plan for VDSL2 on the other hand was to reach 24M customers by 2016.

The German regulator was looking for “procedures to use vectoring even without an exclusive right to the control box."  Alcatel and Deutsche Telkom initially claimed that multi-DSLAM vectoring was impossible and that DT’s “Highlander” network would have to be exclusive.  So the regulator went shopping and apparently settled on Adtran who positioned its product as compatible with a multi-provider system.  My suspicion is that Ikanos is the only chipset maker truly capable of this although it is rumored that Lantiq may be supplying some of the chips despite their offering’s inability to communicate cross-DSLAM.  Even if this rumor is true, IKAN’s CEO has confirmed that this does not preclude Ikanos from also supplying chips or even from Adtran dropping Lantiq after comparing both solutions in the field.  What makes Adtran's win over ALU even more surprising is that Adtran had no public trials or demonstrations announced versus Alcatel who had been involved in a dozen well publicized trials and was the incumbent provider for the network.  Again this is speculation as Adtran hasn't publicly announced their chipset provider but their win over Alcatel-Lucent points to cross DSLAM vectoring or NodeScale Vectoring as a very powerful differentiator.

Huawei, the #2 DSLAM vendor after Alcatel, won a contract to supply equipment for Fastweb's roll out of 3.5M lines that will need to compete/cooperate on Telecom Italia's network.  It's understood that Huawei will be using Ikanos chips proving once again the value of NSV over the competition especially considering Lantiq and Huawei's close relationship dating back to 2002 and the fact that Lantiq was just recently awarded Huawei's Supplier Award for 2012. 

Huawei developed the "World's First Node Level Vectoring Prototype" back in late 2011 and stated that "VDSL2 performance will be greatly enhanced only when NLV is used on all cables connected to a site. By comparison, the board level vectoring (BLV) technology and system level vectoring (SLV) technology cannot achieve such performance enhancement."  Recent comments by Han Yufa, general manager of the Access Network Product Line, on the unveiling of IKAN's Velocity 3 chipset have confirmed that Huawei was indeed working with Ikanos back then: “Huawei’s target is to protect the customer’s investment by eliminating the crosstalk of the existing copper plant using a flexible and future-proof architecture. We have been working with Ikanos for the past two years and now we are glad to see Velocity-3′s achievement on its way.”

Catalysts

Australia's National Broadband Network

Australia's recent change in government will very likely lead to another large opportunity for VDSL2 and FTTN deployments.  The new Minister for Communications, Malcolm Turnbull, has been a vocal critic of the National Broadband Network's original A$39 billion plan to lay fiber all over the country.  He has promised to slash the costs of the project by investing in more FTTC and VDSL deployments.  The NBN which was set up by the outgoing government is more than 2 years behind its original schedule for fiber rollouts.  ASSIA has applauded this decision " to use vectored DSL to bring 100 Mbps quickly to nearly all Australians...Broadband technology has advanced considerably since the National Broadband Network plan was conceived in 2008, and DSL offers a powerful alternative that can actually bring more bandwidth into a neighborhood than GPON fiber and at 10 percent to 20 percent of the cost.”

CEO's Amended Comp package

Tallwood Venture Capital and Dado Banato have been involved for over 4 years now and Dado hasn't tolerated underperformance at the CEO position as evidenced by his quick dismissal of Michael Gulett after seizing control of Ikanos and the less than one year tenure of John Quigley who Dado hand-picked due to his previously successful experience with him at SiRF Technologies.  The fact that the board was willing to amend Omid Tahernia comp package to allow for a full acceleration of the vesting of all equity grants in the event of a change of control suggests that Dado is satisfied with his performance so far and/or is seeking to incentivize him to cooperate in a possible sale sometime down the line.  I'm sure Tallwood's initial plan was always to sell Ikanos as it truly is a minnow compared to its peers but the timing of this event prior to large scale 2014 roll outs and likely design win announcements is interesting.

CTO's recent insider purchases

As previously discussed, this vote of confidence from an insider with better technical knowledge of the product than any analyst could ever hope to have, may convince new investors or previously burned ones to revisit this story.  Again, never in the history of this company has a C-level executive bought shares in the open market (excluding Omid Tahernia's token purchases in 2012 shortly after being appointed CEO which were at less than a dollar and actually perfectly timed as shares proceeded to double over the next 6 months).

Positive revenue growth and any further design win announcements

Management guided to 3Q13 as trough revenues so any sustained period of revenue growth should help and obviously any major announcements or contract wins which are expected in the next few months should gap the stock higher.

Valuation

I've intentionally left valuation to be discussed at the end as I think understanding the history, market opportunity and competitive dynamics are crucial to the thesis as Ikanos is obviously cheap on an EV/Revenue basis if one can get comfortable that revenues won't continue their decline.  There are also no direct comps as Lantiq is private and Broadcom is an entirely different beast.  So this will be a very crude exercise without any fancy valuation models.

Another subscale fabless semiconductor company that VIC members may be familiar with from YCOMBINATOR's fantastic write up is Mindspeed Technologies.  MSPD is the only publicly listed competitor on IKAN's recent prospectus that is a somewhat relevant comp considering it also makes chips for broadband related equipment although its end markets are primarily FTTH and wireless infrastructures.  The recent acquisition of MSPD by MACOM conveniently provides a rough idea of the lower range for IKAN's take out value considering IKAN just reported trough revenues in 3Q13.  MSPD was acquired for $246M net of cash which was approximately ~1.6x TTM revenues.  Or one could look at it from an R&D perspective and see that MACOM roughly valued MSPD at a slight premium to its last 4 years of R&D spend. 

On those two metrics, IKAN is currently worth anywhere from $153M to $242M in EV or $1.93 - $2.85 per share or 54%-128% higher than current prices.

However, as I've tried to explain, IKAN is in the midst of a transition from declining end-of-life ADSL products to rapidly growing next generation VDSL vectoring products.  In 2010, VDSL accounted for approximately 54% of A/VDSL revenue and by 2012 had increased to ~76% - management hasn't provided more recent numbers but it's no doubt even higher now as most of the significant revenue declines have been attributed to ADSL roll off.  The Linley Group noted that VDSL CPE Silicon revenue exceeded ADSL revenue for the first time in 2012.

The near term (next few quarters) revenue growth opportunity is in the accelerated ramp of their CPE revenues which as of 3Q13 comprised 78% of revenues.  IKAN currently has 19 confirmed design wins for their Fusiv CPEs with about half going into production in the next 6-9 months.  And although this write up didn't focus at all on the well-known growth opportunity in fiber/LTE access technologies, IKAN's CPE business does actually serve those end markets as their Fusiv family has 2 main chipsets one optimized for VDSL (Vx18x) and the other for fiber/LTE/other broadband tech (Vx17x). 

2014 should see the majority of these design wins ramping production as well as several more announced design wins (at least 10-13 more in the pipeline).  Q4 revenue guidance was only 17-19M due to continued slowdown in their legacy business but this was to be expected as management has said repeatedly that Q4 would be the inflection.  The slight uptick in revenues from Q3 to Q4 guidance would suggest that VDSL growth is finally beginning to overtake ADSL declines and again point to 2014 as the real opportunity.

The CO Velocity-3 opportunity is also slated for 2014 with their design win in April going into production in the back half of 2014.  Management rhetoric on when to expect additional design wins has gone from "There's still a fair amount of work ahead of us before additional design wins are secured" in Q1 to "we anticipate these engagements to result in additional central office, CO, design wins in the coming months" in Q2 to "we're making very good progress in our engagements with other major OEMs and expect to close on additional Velocity-3 design wins in the next few months" in Q3.  CEO Omid Tahernia reaffirmed this in the Q&A: "we're very specific about sockets that we ultimately proclaim as a win. So I think a number of those are expected in the next few months."

This is significant as the CO business represents the NodeScale Vectoring competitive advantage discussed previously, carries higher margins than the CPE business, and is projected by the Linley Group to more than double by 2017.

Specifically, the Linley Group projects annual global IC market revenue from the sale of VDSL CO solutions to increase from less than $150M in 2012 to nearly $300M in 2017.  VDSL CPE solutions is projected to increase from about $275 million in 2012 to more than $350 million in 2017.  The global IC market revenue for VDSL CO and CPE solutions is therefore expected to increase from ~$400M in 2012 to over $650M in 2017.

Keep in mind this is not the entire addressable market for IKAN as IKAN has in its corporate presentations referenced a $1.8Bn addressable market which it estimates to include all broadband technologies over copper.  However, even using Linley's much more conservative projections yields a 10.2% CAGR in their market and is quite significant in the context of a broader industry that has suffered double digit declines over the past few years. 

IKAN current valuation is giving no credit to this opportunity due to double digit revenues declines as well as significant market share erosion.  CEO Tahernia claimed on the Q2 call that "from an installed perspective in VDSL, we're in the 50% market share range today."  However in 2012, from a shipment perspective, IKAN had only ~24% of the $400M VDSL CO & CPE market.  We know that IKAN did not begin selling its Fusiv CPEs until Q3 2012 and we also know over half of their design wins have not reached production and even the ones that have recently will take a couple quarters to reach peak production levels so by and large these wins have not yet meaningfully contributed to revenue or market share gains.  We also know their flagship Velocity-3 CO solution won't begin shipping till late 2014.

So if IKAN had 24% market share with the majority of its next gen products not yet in production, I think it would be conservative to estimate they could reclaim at least 1/3 of the market once they all ramp.  This would equate to nearly $200M in revenue by 2015 just from VDSL - excluding ADSL, PON/LTE , and other (WiFi/Ethernet) related revenues.  A 1.5-2x EV/Revenue multiple at that point would have IKAN at well over $3.  If NodeScale Vectoring allows IKAN to regain their historical 50%+ market share then this obviously works out even better.

At the today's EV of ~$87M, the market is completely overlooking this emerging growth story.  Investors are able to participate in several years of R&D at a fraction of the cost and which will soon pay off with nearly 20 confirmed design wins across their CPE and CO solutions and several more to be announced in the next few months.  Recent events at the C-level such as insider buying by the CTO and an amendment to the CEO's compensation package suggest insiders see significant opportunity in 2014 and beyond as well.

Risks

The largest risk is more carrier delays in deployment although for all the reasons discussed above the environment seems primed for growth.  Another risk is that Ikanos will keep burning cash on R&D without the revenue growth to support it although this is unlikely as they've spent the last 4 or 5 years on this current product portfolio which has proven successful already with nearly 20 design wins on their way to full production and deployment.  These design wins should also mitigate the high customer concentration as three customers accounted for 49% of sales in the first nine months of 2013.  There is also the risk of another poorly timed and dilutive secondary but the most recent offering provides a good buffer and should provide the necessary time to bridge IKAN into their anticipated revenue ramp.  Looking at IKAN's history, one might also fear more management turnover although as discussed in the catalyst section the recent amendments to Omid's contact would seem to mitigate this fear as well.  Ikanos could be unable to compete effectively against its larger and more well financed competitors but recent wins suggest that being subscale hasn't affected their offering’s attractiveness to vendors/carriers. 

Catalyst