Reflections From The Road (Part II)

By Edward Schneider :

Last year's Reflections From the Road write-up focused solely on Cloud Computing technology sectors (Big Data, M2M, and SaaS). The ensuing piece covers many road-trips, technologies, and company notes that I will cover in chronological order. For those readers who prefer company reports including fast-growing pre-IPO candidates - focus more on the ODDO Midcap Forum and Needham Growth Conference section. For those readers who prefer technology trends with a few highlighted companies, focus on the other sections.

December 2013 - Silicon Valley

Carrier Network Virtualization Conference in Palo Alto - see here for the complete report.

Pivotal Systems visit - private Quan portfolio company providing leading-edge gas flow controllers to advanced semiconductor fabs.

Procera Networks ( PKT ) visit - will discuss in more detail later in this article.

VC Task Force event - several seed-stage startups presented, manly software and web service business models. Nothing disruptive, mainly evolutions of existing business models in new verticals. For example, one high-scoring start-up was U-surance, which would charge insurance agents $10/month via a mobile app for relevant newsfeeds and provide real-time answers to client questions.

One takeaway is the impressive electronics and software ecosystems in Silicon Valley that evolve and become ever-more encompassing. On the east side of San Jose, Samsung ( SSNLF ), Cisco ( CSCO ), Applied Materials ( AMAT ), Lam Research ( LRCX ), among others, constructed enormous fabs and research centers within a few miles of each other, with Intel (INTC) close by in Santa Clara. Nearby semiconductor startups have the advantages of customer/supplier proximity and attracting top industry talent. Similarly, San Francisco and Mountain View/Cupertino have developed vibrant Web software ecosystems. The level of collaboration and brainpower at the Googleplex (GOOG) was especially impressive.

January 2014 - Lyon, France

ODDO Midcap Forum - a large gathering of small-to-mid-cap French companies. Met with Parrot (PARRO.PA), Inside Secure (INSD.PA), Riber (RIB.PA) , and Rentabiliweb (BIL.BR). The stock prices of both Inside Secure and Rentabiliweb have doubled from depressed levels since my early-January meetings, as they are exiting transitions with newer business models. Rentabiliweb now has an online credit card processing platform for the French market (in addition to its legacy web dating, astrology, and other web service offerings). Inside Secure moved towards software-based security for mobile applications, and away from the capital-intensive contactless chip market.

Parrot is a highly innovative company (see this 2012 article ) that is having a difficult 2014 as its legacy automobile hands-free voice and audio systems decline, while its next-generation automobile infotainment offering is being seeded in the market. In the meantime, Parrot has a growing consumer drone business, which is now moving into the commercial market. Riber's molecular beam epitaxy equipment enables the production of high-speed chips (e.g. for data centers), thin-film solar substrates, and OLED flat screens. The company, however, is a few steps removed from these interesting end markets, leading to small, lumpy quarterly sales, with a particularly weak Q1 2014.

January 2014 - New York

Needham Growth Conference - the premier small-cap growth conference with over 400 quality public and private companies. Below are notes from that time on particularly interesting private and public companies that I met.

Select Private Companies Notes from Needham Growth Conference as of January 2014

These private companies fall into three groups: A) hyper-growth/hot-space/pre-IPO - 1) Actifio, 2), 3) Logi Analytics, 4) LogRhythm, 5) Nutanix, 6) Outbrain, 7) Roku, and 8) Zscaler; B) strong growth/impressive-semiconductor-technology/pre-IPO - 1) Impinj and 2) Tilera; and C) rebounding/raising-money-now/cheaper-valuation - 1) N-trig and 2) Tagged.


Disruptive copy data management software provides single-copy back-up and instant access to all data, as opposed to several back-ups/data protection silos and related applications used today.

Copy Data Virtualization

Competitive advantages: dramatic cost savings, decrease in computing power requirements, faster data access, single source of information. Management claims 10x operational impact via: 100x faster restore, 90% decrease in operational processes, 130% product storage performance gain, 95% lower storage cost, single backup, no storage vendor lock-in. Suppose to be easy to implement, and works with all storage devices.

Compete with EMC (EMC), Symantec (SYMC), and Commvault (CVLT), among others. Beat-out EMC at airline reservation mandate. EMC bid $10M. Actifio won at $2.3M with 80% GM and a more robust offering than EMC. Target market is 5 TB or more of storage, mainly large to mid-size enterprises (not SMEs), equates to a TAM of $46B.

Revenues, by my estimate, are over $50M, half are international. Actifio has hundreds of end users, and partners with service providers like Fujitsu (FJTSY) and Dell among others. Actifio uses a SaaS model, charging $0.44/GB for data protection plus $0.05 to $0.10/GB (and growing) for additional services. Bookings were nominal in 2011, and then rose sharply in 2012 and 2013.

Based in Waltham, MA near Boston, Actifio has 260 employees, 34 granted patents, and took in $107M in capital thus far.

Industry-leader in the emerging connected-home solutions market with over 1M subscribers. offers various services such as thermostat control, video monitoring, alarm settings, etc., with real-time alerts and 24/7 emergency response. Users can remotely monitor these services via a mobile app.

Sensors are installed throughout the home, with a dedicated two-way connection to the operations center. An alarm is then routed back to the user via SMS and mobile app. The company supports many stand-alone devices and applications, using several wireless frequencies (Z-Wave, Zigbee, 900 MHz/UHF, etc.), in an effort to be as open as possible. partners with over 4,000 dealers to market to customers, install the service, and provide emergency response. The dealers charge $40 - $60 per month, with taking $6 per month on average. has over $100M in revenues, and is growing rapidly. Three-quarters of revenues are recurring subscriptions (the remaining is hardware and other), creating a scalable and profitable business model. This company is currently cash-flow positive and profitable.'s success has not gone un-noticed. Recently, AT&T (T) and Comcast (CMCSA) have been trying to enter the market. But has a strong foothold with its extensive dealer network in this large and growing market. taps into some very hot investment themes - Internet of Things and Connected Home, which combined with its leadership position in mobile-based home security, fast-growth, and high cash-flow generation, should lead to a premium valuation. is mostly self-funded. The company had a small Series B round last year, partly secondary shares for early investors and employees to take some money off the table. The company is based in Vienna, VA near Washington DC.


Impinj is the dominant market leader in passive RFID tags, readers and software. They have become the low-cost producer of UHF RFID chips/antennas, with the best-performing features: read rates: 1000 tags/second, two-way range: >30 feet, no battery, rewriteable, and no line of site needed. Impinj has a lot of proprietary non-volatile memory technology in their chip design with 150 issued patents.

I know the CEO, Bill Colleran, personally. Bill has participated on RFID industry panels that I organized, despite the fact that Quan was an investor for part of that time in a competitor - Alien Technology. Bill's business savvy and unique expertise in engineering, legal, and investments, enabled his prior successful start-up to be acquired by Broadcom (BRCM), and subsequently led Impinj to replace Alien as the passive RFID industry leader. In short, Bill is a top-notch CEO.

RFID went through the peak of the hype cycle from 2004-2006, before coming down to the reality of actually building-out the RFID ecosystem thereafter. Today, RFID is finally taking off. Standards have been agreed and established, and commercial usage is on the rise. Retail is the largest demand driver; RFID provides accurate real-time inventory snapshots, with labor-saving automated RFID scans. Impinj's embedded RFID chips also offer security and authentication for consumer/electronics, OTC and prescription drugs, transport, and other industries.

Revenues are growing 25% to 30% annually in the $50M-to-$100M range. Gross margins are above 50%. The company is profitable and cash-flow-positive with a scalable business model, as incremental revenues fall to the bottom line. Nearly 4B RFID tags were sold industry-wide in 2013 at around $0.05 each. As RFID tag volumes mount, prices decline, which increases adoption more, and leads to even more tag volume increases. As the low-cost producer, Impinj's market share, revenues and profitability will benefit from this virtuous cycle. Bill said it will be interesting to see Impinj in a few years when industry volumes are in the tens of billions (it may take five years in my opinion).

Founded in 2000 and based in Seattle, Washington, Impinj's several investors include ARCH, Polaris, and Intel.

Logi Analytics

Logi Analytics provides easy-to-use data analytics, and delivers that data to a multitude of end users. Management's vision is "Analytics Everywhere".

Data Delivery

The company utilizes dashboards and reports with data analytics and tables. Differentiation lies in its web architecture that took a number of years to refine, and its built-in collaboration which makes Logi Analytics easy to share with others.

Logi Analytics competes with QLIK Technologies (QLIK), Tableau Software (DATA), IBM (IBM), Microsoft (MSFT), and others. In Gartner's Data Analytics Magic Quadrant analysis, Logi Analytics was recognized for the first time as a Challenger.

The company uses a B2B model to serve 1200 clients presently across many verticals, with particular traction in the health care market. Management did not provide revenue figures, but stated that 80% of sales are recurring, and that revenues grew 600% over the last five years. Assuming a low base, I would estimate revenues of $30M to $50M. The company has 450 OEM distributors today, versus only 30 a few years ago.

There is not a real proprietary advantage to this business. The company's success so far has been driven by a well-designed product filling an unmet need for data analytics in the entire organization (as opposed to CXOs), and judging by the CEO's enthusiasm - 200 passionate Logi Analytics employees.

Based in McLean, VA near Washington DC, Logi Analytics raised $50M in capital, including $25M in the autumn of 2013.


LogRhythm is one of the leaders in SIEM - Security Information & Event Management, as ascertained by Gartner. SIEM is a $1.6B market, and is part of the larger $5.9B Security Intelligence market. LogRhythm basically prevents security breaches. Specifically, the company 1) detects and responds to security threats and breaches, 2) monitors critical digital assets, and most importantly, 3) delivers early identification and fast response to IT incidents.

LogRhythm's competitive advantage over peers like McAfee/Intel and IBM is its advanced analytics to detect attacks while they are occurring before the malware becomes embedded. LogRhythm full SIEM solution includes 1) Log Management, 2) File Integration and Monitoring, and 3) Real-time Host and Network Forensics.

LogRhythm has a diverse base of over 1500 customers, with 100 new customers added in Q4 2013. Revenues and bookings exhibited consistent 40% annual growth from 2010 to 2013, and management expects top-line growth to accelerate beyond 40% per annum from 2014 to 2016. I estimate current revenues to be in the $50M to $75M range. LogRhythm's 40% historical growth rate nearly doubles the 22% rate experienced by the SIEM industry (which coincides with management's claimed 55% win-rate versus competitors). LogRhythm also hangs onto its customers with a 95% maintenance renewal rate. Revenue visibility is high with large recurring revenues from support and maintenance.

Based in Boulder, CO, the company was founded in 2003 and has 240 employees.


N-Trig makes integrated multi-touch and active pen solutions for mobile and desktop displays. N-Trig's patented technology (50 granted and 75 pending, won past patent challenges) provides a natural pen-on-paper feel, while making broader mobile touch experience more collaborative and productive.

N-Trig more or less owns this emerging pen-and-touch market, which is taking market share away from the broader touch-only market. Specific competitive advantages include 1) can do verified signatures on documents (cheaper than other solutions), 2) simple single sensor/touch technology has inherently lower cost and more flexibility, 3) the pen is device-agnostic, and 4) N-Trig's pen is much more advanced than static pens with sponges attached (a.k.a. pogo sticks). The ice-breaker, however, is its soon-to-be-released rechargeable pen (on-device, 15-second charge enables 4-hour battery life), which should increase adoption.

N-Trig was founded in 1999, transitioned from a games-with-touch technology to its current pen and capacitive touch business in 2007, released its first battery powered pen in 2009, and shipped 4M pen/multi-touch systems to date, including 56 design wins with major consumer electronic vendors. In 2012, however, N-Trig experienced a hiccup, when HTC's (2498.TW) $40M in 2011 revenues disappeared and Windows 8 was delayed, leading to a sharp fall in 2012 revenues.

§ Breakeven is expected to be reached in the second-half of 2015, * based on 50% of 2014 forecast

The rise in the gross margin since 2011 is due to the gradual elimination of incorporating the glass display in its product. Glass was 50% of 2011 revenues, but only carried 12% margins. Glass also weighed down 2012 gross margins. In 2013, new partners like Asus (ASUUY), Sony (SNE), Acer (2353.TW), etc. came on board, which led to a sales rebound. The momentum is carrying into 2014 according to management.

The go-to-market strategy is 1) continue to target traditional OEMs with the chip-in-pen/SOC-in- device solution with a greater number of apps for the pen, 2) target pen retailers like Pilot and Parker, and 3) expand price point offerings based on quality and features.

The big upside lays in a recent collaboration agreement with Intel to sell chips with N-Trig's digital technology. This collaboration will create industry standards for pen and touch displays based on N-Trig's technology. Intel's independent study concluded that 78% of people prefer a stylus in their next device. Even Apple (AAPL) is talking about an iPen, according to management.

Management is also targeting Samsung, which uses Wacom's static pogo stick pen that has the added cost of partnering with a haptic vendor (versus N-Trig's integrated simple solution). If Intel and Samsung sign-on as management expects, this will create $300M to $500M in revenues for N-Trig.

Based in Israel, N-Trig has 130 employees. The company raised capital of $120M since inception from major VCs like Canaan and Evergreen, as well as Microsoft.


Nutanix offers virtualized data storage that eliminates the cost and complexity of storage area networks (SANs), while increasing speed and configurability of Virtual Machines (VMs). Through a localized Nutanix Controller on each VM in the datacenter, the data path can be localized, as opposed to having to traverse the network to an external SAN platform.

Nutanix brings software to the blades in the customers' storage system, and better utilizes the flash memory in the server. This multi-purpose box - processing, storage, and networking - replaces several specialized servers and applications. The large cost savings, speed acceleration, and scale and design improvements have created a burgeoning demand from both large and mid-sized companies. Today's small TAM of $300M is expected to jump to $10B in five years. Revenues are currently $80M and are increasing 40% quarter-over-quarter.

Nutanix just closed on a $101M round at a $1B valuation driven by the hot storage virtualization space and Nutanix explosive growth. Santa Clara, CA-based Nutanix raised $173M in capital since its inception merely four years ago.


Outbrain's online content discovery software is channeled into two businesses. First, Outbrain is used by 2500 blue chip customers' web sites to engage their audience in content that is interesting to them, while driving their audience to higher-value/more monetizable web pages. Second, Outbrain's 40 proprietary recommendation algorithms help find the right audience on Outbrain's network of over 300 premium web publishers.

The usage numbers (60B impressions/15B page views monthly) and click-through rates (1.48% for Outbrain-supported display ads versus industry average of only 0.11%) are impressive. Revenues have doubled in each of the last four years, and were over $100M in 2013. This revenue growth is driven by a strong and visible ROI to its customers. The revenue split is 80% US and 20% overseas.

I met the CEO at the Needham conference one year ago, and he delivered on his projection to double revenues. This is a very talented team that worked together at a prior successful online media start-up (Quigo) which was sold to AOL (AOL). They had a vision to build a web recommendation engine with links that consumers can trust to find what they want. Now they are monetizing that vision.

The company was founded in 2006, has over 300 employees, with HQ in New York, NY and R&D in Israel. They have raised capital of $100M, with the latest round of $35M coming in October 2013. Among the investors are Geneva-based Index Ventures.


Roku provides video streaming content to 8M consumers utilizing its own video-streaming players. Roku offers 1200 channels versus 27 from Apple and 14 from Chromecast. Roku is an open platform, as opposed to Apple's closed platform. The quality and breadth of the content is impressive - 31k movies, all sports ex NFL, music including Pandora (P) and Spotify, weather, family and kids, workouts, etc. The main prime content missing from Roku's offering today are ESPN and CNN which cannot be unbundled from traditional offerings (Roku does not yet offer bundled content).

Roku video-streaming players sell for $49 and $99. Roku also has a video streaming stick that sells for $69 that just plugs into the MTL TV port. In 2014, the company will launch Roku TV manufactured by TCL (TCCLF) and HiSense (HISEF) with Roku's operating system embedded in the TVs. This will provide even easier and faster Roku TV access (simple UI, fast search, convenient remote control) and will accelerate Roku subscriber growth even more. Roku anticipates other TV manufacturers will follow.

Roku has three revenue sources: 1) sale of HW devices - currently the largest revenue contributor with GMs of only 20%, 2) revenue-sharing of platform services from third parties - advertising, games, etc. - fastest growing revenue source, and 3) licenses and fees related to Roku's operating system.

2013 revenues increased substantially to over $180M, driven by 3M new subscribers in 2013 alone. Roku's subscriber base has now hit the critical-mass tipping point for Roku to accelerate high-margin recurring fee growth. The Roku TV initiative will accelerate this virtuous recurring revenue cycle even more.

Roku was founded in 2002 by Anthony Wood {CEO}, the inventor of the DVR. Roku is based in Saratoga, CA. I met with Anthony Wood, CEO last year, and this year met with Erik Bardman, CFO (recruited in 2013 from Logitech (LOGI)). Roku has taken in $130M in capital to date, including $65M in May 2013.


Tagged is a social discovery website built for users to meet new users through games, common interests, referrals, browsing profiles, etc. The company's user base has grown exponentially via viral marketing and mobile (250% growth in active mobile users last year). Tagged now has 300M registered members, with 11M monthly active users in 200 countries (only 26% of active users are based in North America). Tagged's success has been helped by its data science expertise (6 patents and 41 patent applications).

The key success driver is matching-up users. Tagged registers 2.4M matches per day or about 1B annually. After a match between users is established, they start messaging - 10B messages were sent in 2013. Thus, Tagged is well-ranked versus other social networks with a #2 engagement ranking, second only to Facebook (FB).

Tagged has successfully monetized its engaged users with virtual currency (60% of 2013 revenues), advertising (37%) and other (3%). 2013 total revenues reached $56M versus $23m in 2009. This growth was achieved with $24M in capital funding since inception in 2004. While Tagged has been profitable since 2008, top-line growth has eased in recent years. Management is now seeking a capital infusion to accelerate future revenue growth.

Management would use the capital to ramp marketing (CRM, branding, etc.), expand the team (organically and via small acquisitions - five done so far), and via the following product launches: 1) newsfeed driving engagement - meeting people based on content will accelerate in 2014, 2) tag prime - comprehensive redesign of website to create unique, customized profile pages, and 3) professional messaging - to monetize messaging in the future.

Management sees Tagged at the intersection of three large, fast-growing markets: Social Networking ($16B TAM), Social Discovery ($2B TAM), and Social Messaging ($9B TAM). Thus, the potential exists for much larger revenues.

Based in San Francisco, CA, Tagged has 175 employees.


Tilera designs high-performing multicore processors for data centers and high-speed networks. Tilera's multicore processors are based on its iMesh technology - scaling throughput at 4x-to-the-square-root-of-n for each additional core - versus less-scalable traditional core designs. This proprietary (107 patents/claims) competitive advantage has standard Tilera chips scaling up to 72 cores (using 64-bit architecture) at networking speeds of up to 80 Gbps today, well ahead of peers. Tilera has designed 144 core processors at even faster networking speeds. Tilera's chips are not only fast with high throughput, but are low power, creating a very high performance per watt.

Data centers are now launching software defined networks that require separate data and control planes. Server processors, however, have not been architected for data planes (leading to too many server processors). Tilera has responded with TILE-IQ product that plugs right into a standard server, and then partitions data and control planes. Tilera's main market today, however, is the broader heterogeneous networking market (ADCs, DPI boxes, firewall servers, adapter cards, etc.) for customers like Cisco, BT (BT), Checkpoint (CKP), etc. The combined specific addressable market for Tilera is $2B today.

I have followed Tilera for a number of years. Commercialization took longer than expected as the market had to catch-up to Tilera's capabilities, and management needed to work through software/design challenges that go hand-in-hand with multicore processors. But in the last few years, the market's need for speed (à la Top Gun), Tilera's successful Linux integration, and the return of co-founder/CEO Devesh Garg and $45M capital infusion in 2011, have helped turn around the company.

One critique that I heard from a potential adapter card partner/competitor is "Tilera must decide if it wants to be a chip company or a networking equipment company. It cannot be both, or it will be competing with customers." Judging from Tilera's recent foray into adapter cards and rack-mounted network boxes, management is running the risk of competing with its customers. The commercial logic of moving up the value chain is evident, but must be done carefully.

When questioned about revenue levels, the CEO responded "revenues are sufficient enough and growing to do an IPO in 12 to 18 months." I would estimate annual revenues to be $30M to $50M. Tilera is based in San Jose, CA with 125 employees.


Zscaler is a cloud-based enterprise security-as-a-service offering. Zscaler's pure cloud security solution requires no additional security hardware or software. Customers no longer need to backhaul traffic into the corporate network, and can instead have direct policy enforcement and traffic analysis with cloud-based proxies. Management labeled Zscaler the Akamai (AKAM) of Security given its speed, security and cloud infrastructure. Gartner concurs with management, positioning Zscaler as a top Leader in its Magic Quadrant analysis.

Zscaler's proprietary storage scalability can quickly condense 7 petabytes of data into 64 gigabits without compromising accessibility to that data. This enables Zscaler to efficiently monitor billions of internet sessions from over 100 data centers around the world.

Zscaler charges annual subscriptions for email, web, mobile and direct internet security. The company has 12M users around the world with large blue-chip clients. According to management, revenues are rising rapidly, gross margins are over 80%, and the company is cash-flow positive. Revenues are $50M+ today, and management is "targeting $100M for a 2015 autumn IPO".

The COO, Lane Bess, called Zscaler a "FireEye (FEYE) without the expensive boxes." He added that FireEye cost $10M in hardware with one customer versus $1M with Zscaler. Main competitors include Blue Coat and Websense.

Lane was previously CEO of high-profile Palo Alto Networks (PANW), before joining ZScaler in 2011 (he is also an advisor to Nutanix). Lane stated that Zscaler has a larger opportunity than PANW.

Based in San Jose, CA, Zscaler was founded in 2008 by CEO, Jay Chaudhry. Most of the shares are held by founders and employees. Remaining shareholders are customers and Lightspeed Venture Partners.

Select Public Companies Notes From Needham Growth Conference as of January 2014

The following public company prospects had attractive businesses and prospects. The three positions that I would initiate were Sequans (SQNS), CEVA (CEVA), and Gowex (GOW.MC, [[LGWXY]]). The other two prospects - Immersion (IMMR) and ChyronHego (CHYR) - I would put on a watch list. Quan Technology Fund had been long Sequans, and may initiate a position in CEVA.


Please refer to previous CEVA articles in May 2014 , February 2014 and February 2014 , which provide a current, in-depth view of CEVA.


ChyronHego provides live video broadcasting with added graphics (Chyron), and related sports tracking technology (Hego). Chyron and Hego merged in early 2013.

The primary merger synergy is transferring Hego's impressive sports-tracking technology (e.g. TRACAB Player Tracking System and Paint Sports Telestrator) from European soccer broadcasters to professional US sports. In a three-hour NFL football match, the actual action on the field lasts only about 11 minutes. This leaves a lot of time to be filled by ChyronHego's sports graphics and data analytics. ChyronHego recently landed Fox Sports including this year's Super Bowl, as well as MLB. The plethora of other US professional sports channels should be next. The bottom-line will also benefit from cutting some fat off US operations including Chryon's previous CEO.

The traditional broadcast graphics products TAM is $250M. The future TAM is about $1.4B when adding virtual advertising ($150M) and broadcast graphic services (>$1B) like online streaming, cloud services, and especially sports clubs' and leagues' use of sports analytics and player tracking. Market growth drivers include 1) rising video consumption (need better/more content, new online streaming of second-tier events), 2) live programming cannot be downloaded, 3) sports and sports data analytics are expanding quickly, and is the low-hanging fruit for ChyronHego today, 4) in-content advertising, and 5) broadcast services outsourcing trend. Management said that the US sports market for content creating videos and sport tracking is worth over $25M (I was hoping for more).

The competitive landscape is highly fragmented. According to management, industry consolidation is likely due to the higher investments required in order to compete in this expanding market.

ChyronHego has $40M in LTM revenues through September 2013, with broadcast graphic products accounting for 64% of sales, and services 36%. Revenue visibility is high, with 50% of 2014 budgeted revenues already in backlog. Revenue growth in 2014-15 should be driven by 1) geographic expansion into Latin America, Asia and Middle-East, 2) product line expansion, and 3) potential acquisitions. Ongoing operating margins are 10% today according to management. The company also has a large net-operating-tax-loss-carry-forward ((NOTL)). While not disclosing any specific numbers, the CEO indicated that future revenues and profits should grow nicely.

Chyron was historically an unprofitable company, and just moved into EBITDA positive territory with LTM EBITDA of $1.9M. Consensus estimates (of only one or two analysts) expect the merged entity to be more profitable in 2014 with revenues of $58M and EBITDA of $6.5M.

*reported LTM EBITDA through Sep 2013 is $1.9M; **2012 & earlier pro-forma results were not available.

The main risk is the nature of the industry, with its historically low operating margins despite generally strong gross margins, and the need for consolidation to cover high fixed costs. Of the three public direct peers, Orad (OHT.F) and Avid (AVID) are currently losing money, and only Vizrt (VIZ.OL, VZRTF) is profitable. Avid is a basket case right now, and Orad is even smaller than ChyronHego. But Vizrt, with revenues of $124M, GMs of 66%, EBITDA margins of 19%, and revenues per employee of $231k, provides a long-term operating target for ChyronHego. Given the CEO's stated intention to grow via acquisition, it would not surprise me if ChyronHego tried to acquire Orad, which is only trading at 0.5x revenues, or a smaller private competitor.

*Avid is in the process of restating financials. Latest figures are only for the 12 months ended September 2012.

Swedish management owns one-third of the company, and there have been five insider purchases (no selling) in the last six months. Salaries are fairly modest. Management and shareholder interests are both tied to the appreciation of CHYR stock.

Currently priced at $2.81, CHYR shares rose 33% in January, after nearly tripling in 2013 from depressed 2012 price levels. At a market cap $86M (and EV of $87M), CHYR is fairly valued at 2.2x LTM revenues and 1.6x NTM revenues, when compared to much slower-growing peer group valuations of 1.1x and 1.0x, respectively. CHYR's EV/NTM EBITDA of 17.2x and price/2014 earnings of 31.2x are expensive both on an absolute basis and relative to its direct public peers. The main justification for CHYR's premium valuation is its higher growth including high-profile wins in the emerging sports graphics and analysis market. But this high growth rate's sustainability is still a question mark.

In sum, my initial enthusiasm for CHYR shares have been tempered by industry dynamics, relative peer group valuation, and the stocks run-up. I also would like to understand the scalability of Hego's service business model. I would not buy CHYR shares at these price levels, but would put the stock on a watch list.

ChyronHego's HQs are in Melville, NY on Long Island.


Gowex (officially called Let's Gowex) offers a "smart city" free local Wi-Fi connectivity platform on streets and public transport that is rapidly expanding in cities around the world (latest count is 81 cities and 3M users). Gowex receives multiple income streams including 1) initial engineering fees after winning the public tender, 2) offloaded traffic revenues, 3) share of roaming incomes from wireless operators, 4) growing number of data/media/geo-localization services and apps (e.g. sale of targeted advertising data, geo-location system mixed with couponing of a nearby merchant, etc.), and 5) recurring smart services (e.g. coin collection for transport, parking apps, etc.).

Gowex protects its revenues and direct customer contact via 1) long-term exclusive municipal contracts that last for 7 years on average, 2) creating new services like WE2 (social neighborhood Wi-Fi) that further entrench and engage Gowex with end users, 3) sharing a portion of recurring revenue streams from certain value-add services with public authorities of different cities, and 4) paying for its own backhaul of network traffic.

Gowex has expanded from just 3 Spanish cities in 2009, to 81 worldwide today including Hong Kong and New York, with a target of 150 by the end of 2015, and 300 by the end of the decade with 100M users. As Gowex approaches its 300-city target, it will become a self-fulfilling global platform with economies of scale and network effect, propelling further penetration (akin to Facebook). For 2014, Gowex has a prospect pipeline of 40 - 50 cities, and should close on 20 - 30 of them in 2014.

Gowex has a very high win rate, losing a total of only four bids last year vis-à-vis SFR in France, two to Arqiva in the UK, and a corrupt local player in a Buenos Aires suburb. Gowex is basically creating its own market model, and does not have any direct competition yet besides two local competitors in Spain and Arqiva in the UK. The main competitive threats come from Facebook (only in San Francisco retailer so far) and Google (only in Starbucks (SBUX) and Kansas City, so far). Gowex's response is to grow as fast as possible to become a self-fulfilling global standard, and to have M2M and Wi-Fi at the home (with services like WE2) in addition to the city. This recently launched home Wi-Fi platform 1) solves connectivity problems (data at home is 30x larger than network capacity), 2) improves social experience for home and nearby merchants, 3) enables monetization via geo-located ads, coupons, etc., and 4) targets SME's advertising dollars à la Google, while giving free Wi-Fi to end users.

Gowex was one of the fastest-growing public companies at the conference. Revenue and EBITDA jumped 3.5x and 8.2x, respectively, over the last three years. In latest reported results covering 2013 1H, revenues rose 40% and EBITDA increased 55%. Spain was 44% of 1H revenues, but that percentage is declining. The fast-growing wireless business rose 50%, and accounted for 84% of 1H sales. The older telecom business declined 3%, and was only 16% of sales. Gowex had EUR 62M ($85M) in cash and EUR 17M ($23M) in debt as of June 30, 2013.

Gowex is based in Madrid, Spain. The company has 180 employees, 70% of which are engineers. I met the founder and CEO Jenaro Garcia Martin and advisor/early investor Paul Ryb. Jenaro has created a startup culture with high innovation and low expenses, and appears to be a focused, hard-working entrepreneur. He alluded to maybe finding a big-organization-type CEO down the road, but not in the near term. Jenaro remains humble and appears unfazed by the $886M net worth of his 52% ownership of Gowex.

Investors have reacted to Gowex's stellar operating performance. Currently priced at EUR 17.40 ($23.49), GOW.MC rose 33% in local terms during January 2014 boosted by signing up a large Chinese city (Ningbo) to a 15-year exclusive contract. GOW.MC jumped 7x since its last secondary financing round in December 2012 at EUR 2.50. Gowex is quoted on the Madrid bourse (GOW.MC) covering two-thirds of trading volume, the Paris bourse (GOW.PA) for the remaining third of trading, and OTC in the US where it trades by appointment.

At its current market cap of EUR 1.26B ($1.70B) or enterprise value of EUR 1.21B ($1.64B), GOW.MC has an EV/2013 revenues multiple of 6.9x, EV/2013 EBITDA multiple of 24.3x, and price/2013 earnings ratio of 41.4x; 2014 multiple of 4.5x, 15.8x, and 26.4x, respectively, are more reasonable. But when factoring in Gowex's 40% long-term growth rate (supported by 57% profit growth in both 2013 and 2014), the stock becomes more interesting. Using a PE/G ratio below 1x as reasonable, Gowex is attractively valued with a PEG ratio of 0.66 based on 2014 earnings.

In the current zero-interest rate/high-liquidity environment, hyper-growth US stocks like Splunk (SPLK) and 3D Systems (DDD) are trading at multiples that at least double those of Gowex. GOW.MC's biggest risk would be a tightening of the US-led liquidity faucet, given the stock's run-up and high absolute valuation levels. The operational risks are more about execution, renewing major contracts that mainly expire in the next decade, and getting ahead of the inevitable competition that is likely to show up. While these operational risks should not be understated, I have seen more onerous risks on average at other technology companies. In sum, I would monitor Gowex shares, take a small position here, and add more on a correction.


Immersion is a leading haptics (i.e. touch) technology company licensing its patents and solutions to the mobile industry (71% of revenues), gaming (16%), medical (7%) and automotive (6%) sectors under the Touch Sense brand name. Immersion's technology recreates physical engagement with electronic devices via a sense of touch. Immersion's core haptic values are: 1) Confirmation - to feel the user interface interaction with touch feedback, 2) Realism - such as immersive gaming, and 3) Rich Communication - such as enabling touch when apart from the screen.

Royalties and licensing account for 98% of revenues, the rest is NRE development payments. Founded in 1993, Immersion has accumulated a massive patent portfolio of about 700 issued patents, and 700 pending patents. Immersion is not afraid to sue infringers, and spent about $5M in 2013 on litigation expenses. In 2013, larger customers like Samsung (35% of revenues) and Sony renegotiated contracts to a fixed licensing fee model, eliminating or reducing royalty per unit upside. Management counters that 1) the smartphone market is maturing, so it is now better to have the certainty of a fixed fee today; and 2) this fixed licensing fee policy increased the number of willing potential customers, leading to the signing of Xiaomi (China's leading smartphone maker) in September. While management's view may make sense at first blush, the elimination of per-unit royalties for the bulk of future revenues reduces scalability and direct correlation with the fast-growing (+25% per annum) haptics market.

To be fair, as part of the Samsung contract renegotiation, the small remaining royalty component will include all shipped smartphone and tablet units, not just the Galaxy product line, as of January 2014. So, near-term these contract renegotiations could be viewed as neutral-to-slightly-positive. Management also mentioned that the fixed-license contracts are mainly annual payments, and only once in the company's history has a license not been renewed. So license payments have been a highly predictable revenue stream.

Immersion's initial market, gaming, has been in a decline over the last few years. Sony's successful PS4 launch in November, however, may reverse the slide, and add to Immersion's revenues (albeit less than before the contract renegotiation, given the higher-fixed licensing component).

While gaming and smartphones may no longer be hyper-growth markets, haptics adoption in those markets is gaining steam, driving a boost in recent revenues. New mobile customers like Sharp (SHCAY), as well as existing customers like Samsung (including Samsung's new smart-watch), are finding more ways to use Immersion's technology. At the same time, Immersion is trying to add touch to new and higher-growth markets like content and media. Haptics can be added to a movie, ad or sporting event on a mobile phone/tablet, and videos like GoPro on YouTube. The emerging wearable computing sector is also fertile ground for Immersion.

Rising revenues in 2013 moved Immersion's scalable business model into the black. 2013 Q3 revenues jumped 59% to $11.3M, with a net profit of $1.2M versus a loss of -$3.0M in the year-ago quarter. Full-year 2013 revenues should be about $48M, a 49% increase over 2012. GAAP net income is estimated to be $6.6M in 2013, versus a loss of -$5.7M in the prior year. Revenues and margins are expected to ramp higher in 2014, with about two-thirds of 2014 incremental revenues falling to the bottom line.

*consensus estimates

Consensus 2015 estimates consists of only one analyst, and projects a decline in profits, despite an increase in revenues to $70M. This is inconsistent with the financial modeling of Immersion prior to 2015, and can be flushed-out with management after Immersion reports Q4 2013 results on February 20.

Currently priced at $11.70, IMMR shares are up 13% in January 2014, after gaining 51% in 2013 and have doubled over the last two years. IMMR shares have a market cap of $335M, and after deducting net cash and equivalents of $65M, have an enterprise value of $270M. IMMR shares are fully valued on an absolute basis using last-twelve months' reported figures, trading at EV/revenues of 6.1x, EV/EBITDA of 38.4x, and a P/E of 85.1x. Based on 2014 figures, the numbers are more reasonable with EV/revenues of 4.4x, EV/EBITDA of 16.5X, and a P/E of 26.6x. IMMR is valued in line with its direct peers. The premium valuations of IMMR and its few public peers are merited in light of their well protected, scalable and profitable royalty/licensing business model. ARM Holdings (ARMH) and CEVA are the most direct comparisons, while Imagination Technologies (IMG.L) had major operational shortfalls in 2013 and has a more levered balance sheet.

Immersion is at the early-stage of a major ramp in revenues and profits. Before purchasing IMMR shares, however, I would need to better understand management's recent shift to fixed licensing revenues, and its effect on the business model. Also, 2015 financial projections should be verified. Immersion also has a large stock-based compensation plan (larger than that of CEVA) for a small company with only 94 employees that could be a topic for discussion.

Based in San Jose, CA, Immersion has 94 employees.


Please refer to previous Sequans articles in May 2014 and February 2014 , which provide a current, in-depth view of SQNS.

February 2014 - Barcelona

Mobile World Congress ((MWC)) in Barcelona is the second largest electronics tradeshow in the world, second only to the Consumer Electronics Show in Las Vegas. One MWC theme was smartphones for the masses. Mozilla teamed up with chip-maker Spreadtrum/Tsinghua so that Firefox OS 2G smartphones can penetrate cost-sensitive markets like India and Indonesia at prices as low as $25 . A second theme was the implementation of LTE (long-term evolution for mobile networks) , and the continued evolution of LTE-Advanced and 5G networks. The LTE build-out has taken longer than many vendors would have hoped, with current deployments mainly relying on Qualcomm's Snapdragon chip-set.

I visited Qualcomm's booth, and received a tour of the Qualcomm Connected Home demonstration, which featured third-party applications developed on Qualcomm's chip platform. The Connected Home is not a new theme, and Qualcomm is not the first chip company to attack this market. Entropic (ENTR) and the Multimedia over Coax Alliance (MoCA) have been targeting the connected home market for several years. But it appears that the timing could be right for the connected home market to finally take off, as related software applications become more pervasive.

I re-examined the mobile health care theme, visiting Fitbit , LifeWatch V , and re-acquainting myself with Aerotel Medical Systems . Fitness bands and apps are in demand. Fitbit's booth was overflowing with people. Strava 's mobile fitness apps have developed a cult following from hard-core sport enthusiasts. LifeWatch V and Aerotel seem to be missing the mobile health-care boat. LifeWatch V requires that each user buy its specially-equipped LifeWatch V phone; not a scalable business model. Aerotel is focusing on the higher-end health care market with bulkier but reliable health monitoring products. At MWC, however, the company launched a mini ECG monitor with a 3G cellular connection. So, management is adapting to the massive consumer mobile health care/wellness opportunity.

Some names that stood-out from the many startups that were visited include Moni Technologies (UK) - mobile web application for money transfers and remittances, EyeSight Mobile Technologies (Israel) - touch-free interactions with digital devices using simple hand gestures, nViso (Switzerland) - uses 3D facial imaging to detect emotions of internet users, and Expway (France) - LTE mobile broadcast solutions.

February - April 2014 - Geneva, Zurich, Paris, Lausanne

I attended several events like Seedstars (finalists from its one-year global tour of start-up companies), LIFT technology conference, ETH Zurich visit, Paris Midcap Forum/company visit, Venturelab EPFL Elevator Pitch competition, among other adventures. One unique political media startup is GovFaces (Switzerland). GovFaces has become a web platform for meaningful dialogues between politicians and their constituents. Recently launched for the EU elections, GovFaces is planning to expand its model for the UK and US elections in the next two years.

On the quoted side, I became incrementally more positive on Weborama (ALWEB.PA) after meeting with management in April. The 2014 Q1 results exhibited the worst decline in recent history, with revenues falling -14%, or -8% excluding the negative Ruble currency translation. The main culprit was the commoditized French media business that is rapidly shrinking from Weborama's revenue stream. But most other parts of the business had double-digit growth ex-currency effects, including revenues outside of France (66% of Q1 revenue) and data analytics. The key will be to make the data platform more automated for users, which management is presently working on. Currently priced at €8.50, the stock is down 66% from its March 2012 peak, and is only valued at 1x revenue. The bad news may be in the price.

April 2014 - Sardinia

Capo Caccia Cognitive Neuromorphic Engineering Workshop - this gathering of PhDs and researchers in biology and computer science/engineering seeking the Holy Grail of mapping and emulating brain functions onto computational systems.

Neuromorphic Cognitive Systems can learn and reason about actions to take in response to the combination of external stimuli, internal states, and behavioral objectives. In layman terms, cognitive machines would function like the brain. The human brain has 100 billion neurons (akin to computational units), 100 trillion synapses (memory), and axons and dendrites (network connectors). We are many years away from fully mapping the human brain onto silicon, including the ability to prioritize various stimuli and objectives. Still, we should not underestimate the amount of progress that we are making today.

According to INI Zurich, two categories of neuromorphic hardware are Simulation and Emulation. Both simulation and emulation produce the same outputs from the same inputs that the original system (i.e. the brain) produces. But simulation does not use the same process as the brain, while emulation attempts to use the same process as the brain to achieve the results. While emulation is more challenging, it has more potential. Simply mapping or simulating the brain without understanding the processing and prioritizing of these neurons and synapses has limitations. I was impressed by the Spiking Neural Network emulation projects presented in Sardinia and Zurich.

That said, both simulation and emulation have their pros and cons. Simulation tends to be more digitally-oriented, can handle very large-scale inputs, process them at very high speeds, but use a lot of power. Emulation is primarily analog, works at lower speeds, but at low power. In fact, emulation and simulation are starting to merge, but it will take a number of years. Also, simulation may work well for more data-intensive data mining applications (e.g. work done by IBM ), while emulation may be more appropriate for neuroscience and cognitive applications.

Deep Thinking Networks can add value to data mining operations. Deep Thinking Networks are different neuron network layers that screen and funnel through massive amounts of data until a final set of results are found. These are not preconfigured screens. The data is treated without any preconceptions. Please refer to this Nature article for more detail.

Besides data mining, other commercial applications exist based on neuromorphic vision sensors. Qualcomm's neuromorphic Zeroth processors can be used in robotics, and in particular image sensors. For example, neuromorphic vision sensors in cars could detect sudden object movement much faster than traditional camera surveillance systems. At the workshop, a presentation was made to enhance dynamic vision sensors with "convolution" for better mapping and tracking of objects. Other commercial applications range from virtual-reality video games, to advanced drones, to helping paralyzed people to move.

A large potential opportunity would be installing some of these self-learning cognitive capabilities on a mobile device. A next-generation Siri that increasingly self-learns your needs and preferences could be an increasingly addictive consumer device.

Please refer to this past article and this interview for more information on neural processing investment opportunities.

May 2014 - London

Network Virtualization & SDN World - this event was the third in a series of next-generation network conferences. As a primer, SDN (Software Defined Networking) is direct programming that tells the silicon switches what to do. An SDN network can be managed as one virtualized switch, instead of managing many switches or routers at one time. This is accomplished by a separate control plane that can route network traffic ideally with one standard protocol (e.g. OpenFlow) as opposed to thousands of hardware routers with their proprietary network protocols. Below the control plane would be a forwarding (a.k.a. data) plane consisting of high-speed, lower-cost switches, as the intelligence rests at the centralized control plane.

Network Virtualization is similar to SDN but with more of a telecom focus that is less tied to the OpenFlow protocol. Network Virtualization focuses on virtualizing the separate network hardware appliances like routers, firewalls, load balancers, and distribution switches that are all over the network into a centralized virtual switch that incorporates all these hardware functions. An Orchestration layer is then needed to create a total solution that can interoperate with OpenStack cloud computing platforms and other open-source and carrier-specific protocols.

However, various master orchestration platforms need to interoperate with one another to make network virtualization a reality outside of any one carrier's network - including the highly-publicized AT&T Domain 2.0 program. The predominant master orchestration systems from vendors such as Alcatel Lucent (ALU) and its CloudBand ecosystem , Cisco, and Cyan (CYNI) currently do not interoperate. A recurring carrier complaint at the conference was the need for standardized APIs to improve interoperability.

A common mantra from carriers like Telefonica (TEF.MC) is the need for openness from vendors. Telefonica took an interesting strategic move by embracing the open-source movement to the point that management has given up control of its application programming projects and entrusted them to the open-source community. These open-source developers are external to Telefonica; the company does not have any internal programmers.

SDN and virtualization really is running on two tracks. The fast track is with Web 2.0 data centers such as eBay (EBAY), Yahoo! (YHOO), Facebook and Google that are deploying today. Most of them are hybrid open-source/proprietary platforms today, although the open-source trend will only increase with time. The data centers have the advantage of fairly homogeneous networks.

The carriers are on the slow track. They have complex, heterogeneous networks with high service requirements for legacy voice traffic. Even the carriers' data center traffic is mostly north-south, as opposed to more internal east-west traffic from the Web 2.0 data centers. Also, carriers have to cross the cultural divide from OSS/BSS protocols to the programming world of SDN and virtualization. Carriers, however, realize SDN and virtualization's benefits of agility/faster time to market, capex and opex savings, no more vendor lock-in, etc. This is why some carriers are starting to deploy SDN and virtualization despite the hassles.

For example, AT&T Domain 2.0 initiative includes a virtual Evolved Packet Core ((EPC)). But virtual EPC will be deployed in parallel to the main EPC platform. Virtual EPC will focus on M2M and enterprise services, configuring the EPC for specific traffic profiles. The bulk of AT&T's EPC production, however, won't be virtualized for a number of years. Maybe the next step would be to virtualize functions such as DPI (deep packet inspection), video optimization, and security firewalls. But this will be a multi-step process that will take several years.

VMware/Nicira (VMW) is emerging as a dominant force in the data centers, providing an end-to-end virtualization solution that includes both open-source and proprietary components. Another beneficiary is Akamai, with its CDN (content delivery network) for caching and accelerating the transport of media content over the internet. Akamai has become the carriers' independent CDN of choice, and with carrier traffic exploding, this puts Akamai in a venerable position.

Cisco is increasingly being relegated to the WAN (wide area networks), after losing some ground to VMware/Nicira in the data centers. AT&T also shook up big vendors like Cisco and Alcatel Lucent by going with startup Affirmed Networks for its virtual EPC initiative. But all is not lost for Cisco, mid-tier operators like PCCW, which do not have the resources of large carriers, are sticking with Cisco and its virtualization path to lower costs. Customer demand for Cisco's virtualization offering is very strong according to Ian Hood, who oversees service provider architectures for Cisco. As for Cisco's main router business, the company is fairly protected for the next few years as Cisco is too high in the food chain in terms of mission-critical core routers and customer dependence. However, Cisco's traditional model of proprietary, robust-yet-costly routers proliferating all over the network is the antithesis of SDN's centralized intelligent switching with fast-but-cheap edge-routers. Over time, Cisco will increasingly be in the situation of cannibalizing its own hardware business with its own SDN solution, rather than losing that business to another SDN vendor. In sum, organic revenue growth will be a challenge for Cisco going forward. It is no accident that Cisco revenues peaked in Fiscal 2013, and are projected to decline 3% in Fiscal 2014.

At $3.76, Cyan's shares have crashed 66% from its May 2013 IPO. Cyan had the misfortune of a substantial order cutback from Windstream (WIN) to less than 10% of Q1 2014 revenues. In 2013, Windstream accounted for 39% of Cyan's revenues. Cyan's skyrocketing sales over the past few years are expected to reverse. 2014 revenues should fall 17% to $97M with an operating loss of -$47M. The good news is that Cyan's orchestration software/Blue Planet partner platform is gaining traction with carriers like Telefonica, Colt, and several others. These hard-earned wins are increasing Cyan's strategic value to potential suitors like Cisco or Alcatel-Lucent. Thus, if Cyan can survive this financial downturn, aided by net cash of $42M in its coffers today, the stock may see brighter days.

The stock prices of independent DPI/network intelligence vendors like Allot (ALLT) and Procera have taken a hit year to date, while Sandvine (SVC.TO) shares have rebounded over that period. The whole group is susceptible to virtualization risk in the coming years. France-based Qosmos has entered the market with virtualized DPI solution. Currently, Qosmos solution cannot scale to meet the high traffic volumes that its DPI peers can handle. Qosmos may or may not be able to scale it traffic capacities in the future; there are signaling and other challenges that it must overcome. But the risk exists. Thus, each DPI vendor needs to have a virtualization roadmap.

Procera's stock has been hit the hardest. Currently priced at $8.89, the stock is trading near the bottom of its $8-to-$16 last-twelve-month range. The stock weakness has been due to a drop-off in revenues from its traditional US cable customers. The cable business bottomed-out at 10% of revenues in Q1 2014. Procera is projecting 15% total revenue growth in 2014, but back-end loaded.

I have spent some time with Procera over the last six months at its HQ in Fremont, CA, at MWC in Barcelona, and at the conference last week in London. The pros and cons of Procera are:

Procera has some talented business development and marketing people, a strong technology team in Sweden, and a vision that has been clearly articulated by the CEO. The company faces substantial risks, but downside is limited at current valuation levels.

Disclosure: I am long SQNS, CEVA. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it. I have no business relationship with any company whose stock is mentioned in this article.

Additional disclosure: I am also long ALWEB.

See also Facebook: Addressing The Growth Slowdown Fears on

The views and opinions expressed herein are the views and opinions of the author and do not necessarily reflect those of Nasdaq, Inc.

The views and opinions expressed herein are the views and opinions of the author and do not necessarily reflect those of Nasdaq, Inc.

More Related Articles

Sign up for Smart Investing to get the latest news, strategies and tips to help you invest smarter.