Frigon highlights tech companies on radio interview

Recently, TFCM's CIO Gerry Frigon was interviewed by our own John Summer on KVEC radio in San Luis Obispo.  Mr. Summer was guest hosting for Dave Congalton as he discussed Taylor Frigon Capital Partners' most impressive technology companies.  You can hear the interview on this podcast, starting at 1:00. http://920kvec.com/podcasts/congalton/05-20-15.mp3
Continue Reading

QuickLogic (QUIK): a venture capital-style opportunity for public market investors (RESEARCH)

QuickLogic (QUIK): a venture capital-style opportunity for public market investors, offering a solution for the incipient flood of sensor data 
05/08/2015

Bottom Line: QUIK is trading at 5-year lows, having been sold off successively in recent weeks for guiding lower than investors expected, and then announcing the departure of their CFO. We believe the company has significant potential, and that the story is not widely nor well understood.

Summary: QuickLogic is an unappreciated semiconductor company positioned in front of a major opportunity in the nascent mobile sensor processing solutions market. In a situation similar to the types of investment opportunities that venture capitalists find interesting, the company is currently valued under $100 million and has a viable solution that addresses a new market the size of which is difficult to determine because it is just now beginning to take shape, but which may be in the billions per year. Of course, in most situations involving a small company with a potentially market-shifting technology, there are plenty of risks, including execution risks; risks that other technologies or approaches might eventually win out; as well as all the risks that come along with the situation of being a tiny company selling into the piranha-tank of the mobile consumer device market dominated by big-name consumer-device OEMs, cut-throat competition and pricing, and fashions and trends that change literally from week to week. (An OEM is an “original equipment manufacturer,” and an ODM is an “original design manufacturer,” and both types of companies are potential QUIK customers. During this report we will simply use the abbreviation “OEM” to refer to QUIK’s customers, who design and/or make mobile devices). 

Unlike a venture-backed start-up, however, the company has a market-proven technology, major customers including Samsung with more big-name customers expressing executive-level interest in their product, revenue run-rates of around $20 million per year or more, over $28 million of cash on the balance sheet, low probability of requiring a future capital raise in order to stay in business (although the company is currently not at break-even yet), and a very experienced management team with decades of industry background in Silicon Valley. 

The stock has been severely hammered in recent months, currently trading around $1.70 and reaching lows below $1.60 in the past twenty-four hours, off of 52-week high of $5.48 in June 2014. On 05/06/2015 the company announced the departure of the CFO, leading to another selloff on 05/07. 

We believe the company management is honest, capable, and has a valid strategic plan which involves the potentially enormous market for sensor data processing solutions in mobile devices containing sensors (a market that is in its early stages now but which will almost certainly be getting much bigger from where it is today), and that the company’s technology addresses a major issue that is on the forefront of every engineering design team in the mobile device business today – power. Further, the company’s silicon has specific architectural advantages which makes their solutions unique and difficult to replicate by competitors. Their solutions are programmable, which is another major advantage, and the company has developed a library of software algorithms which offer added value to OEMs seeking to differentiate their products in the fast-moving and highly-competitive markets for smartphones, tablets, and now wearables.

As the discussion below should make clear, it requires a fair amount of rather arcane technical discussion to understand some of the important aspects of the company’s technology and its advantages in the market that is now developing. While the discussion below is fairly lengthy and somewhat technical, it is only a very simplified version of the big picture. We do not believe that some of the more granular details of the QUIK investment thesis are well understood: in fact, we believe that the company is widely misunderstood, as well as being simply too small to notice for many institutional investment managers and large sell-side research departments.  However, we believe that the company’s technology has significant advantages which are well positioned for an upcoming market in “always aware,” sensor-rich devices, and that the companies which make these devices will likely have to consider QUIK’s solutions because of the power-efficiency advantages that QUIK offers, as well as the flexibility, customizability, time-to-market, software algorithm, and product-roadmap advantages that are equally important but also not obvious without some familiarity with the industry specifics.

Please refer to IMPORTANT DISCLOSURE INFORMATION at end of this research.

Basic Description: QuickLogic is a fabless semiconductor company that designs programmable semiconductors and the associated software that OEMs who use their semiconductors need in order to incorporate those semiconductors into their products. The company produces several product lines, almost all of which are characterized by the use of programmable logic (similar to an FPGA but with some important technical differences), and some of which also incorporate certain hardened logic blocks which are not programmable but which perform pre-selected tasks using hardware for greater power efficiency. Additionally, the company designs software algorithms which function as drivers for the hardware, and which in some of their recent products also offer OEMs a selection of algorithms to perform important consumer-facing functions such as voice-recognition applications, gesture-recognition applications, navigation applications, as well as algorithms for future capabilities that few or no mobile devices and wearables on the market today are even capable of offering. The company outsources the fabrication of their wafers to manufacturers such as TSMC, eSilicon, Globalfoundries, and TowerJazz (JAZZ).

Selected income statement items for the past nine quarters are shown below:



Revenues by geography for the most recent quarter were 62% Asia-Pacific, 30% North America, and 8% Europe.


Business Opportunity and Investment Thesis: Founded in 1988 and headquartered in Sunnyvale, California, QuickLogic pioneered a type of programmable silicon in the family of programmable logic that also includes the more-familiar FPGAs and PLDs marketed by XLNX, ALTR and LSCC, but Quick’s approach had its own set of specific strengths and features including certain power-efficiency advantages. For many years, especially before mobile markets made low power-consumption the kind of critically-important feature that it is today, QUIK mainly sold into military and aerospace markets that needed certain other features that their type of programmability could offer, but outside of these markets QUIK’s advantages were not that critical, and although the company’s stock did run up with other technology companies during the late 1990s and early part of the 2000s, its “legacy” applications did not offer particularly high growth potential.  

More recently, however, QUIK’s programmability and power-efficiency has become much more interesting, as the market for mobile connected devices which did not even exist until about ten years ago has made some of their unique features more attractive to potential customers. QUIK has applied these strengths to the mobile market to offer bridging and connectivity solutions in phones and tablets. 

Bridging solutions refer to the use of an intermediary processor to act as a “display bridge” between an apps processor (the central processor in a phone or a tablet) and the display, almost like a translator between two people speaking different languages. Apps processors are often mass produced with a certain type of display output that does not match the type of input needed by the silicon that is governing the screen display. A display bridge can be a much more practical solution than trying to find (or design from scratch) an apps processor that “speaks the desired language” of the display that the OEM wants to use.

Other common mobile market uses for programmable logic of the type that QUIK offers include solutions for OEMs needing a rapidly-deliverable “fix” or “workaround” for a product going to market. These types of product applications will probably always be a steady source of some revenues for QUIK, but in each of the situations just described, it is likely that the “bridging” or other connectivity “fix”  may be designed-out of future generations of the same product (when the functionality is absorbed into the apps processor, for example). When tablet OEMs including Samsung began using QUIK silicon for connectivity and/or display-bridging roles in previous quarters, revenues spiked, investors began to take notice, and the stock ran up to its 52-week highs (following the record-high quarterly revenues of $11.16 million in the March quarter of 2014). As these more “short-lived” uses for their programmable solutions were designed-out of later generations of some of their initial mobile customer products, however, and as tablet sales globally hit a major decline in the second half of 2014, QUIK’s stock price underwent a similar collapse.

However, while these “shorter-lived” mobile applications for their silicon’s unique advantages have been the main focus for momentum investors in the past 52 weeks (whose interest has also been quite temporary), and while there will probably always be an ongoing market for the kinds of “short-lived” workaround solutions that QUIK’s power-efficient programmable logic can offer to OEMs with these kinds of product needs, the management team of the company has quietly been steering towards a new opportunity that is still in its very early stages: extremely low-power, customizable sensor data processing solutions for the fast-approaching flood of sensor-rich data that is on its way in the near future. In this regard, QUIK’s solutions have the potential to help engineers and product design teams at the OEMs to address two dilemmas which are already problematic but which are likely to become even more pressing as data-processing demands on mobile devices increase.

As devices perform more tasks, processors must execute more duty cycles, and they use electrical power in order to do so, draining the battery and generating heat in the process. There are “hard limits” to how much heat a mobile device can generate before it becomes a problem (you don’t want to burn your customer’s face, for example), and there are also hard limits to how long a battery will last as the processors in the device are called upon to perform more and more duty cycles (after a certain point, the battery will have to be recharged). Engineers and architects are faced with conflicting demands: the device is expected to do more delightful and amazing applications, powered by a smaller and smaller battery that holds less and less charge (the Apple Watch, for instance, appears to contain a 205 mAh battery according to this teardown from iFixit), while users want to go longer and longer between charging periods. This conflicting set of demands poses something of a dilemma for OEMs.

Those demands (driven in part by the desires of the owner of a mobile device such as a smartphone, tablet or wearable, and in part by the engineering realities of amps, battery capacity, and processor duty-cycles) are fairly straightforward and have been part of the picture at consumer device companies for some time, although there are important reasons (hint: sensors) why the processor demands in the first part of the equation may be getting ready to increase dramatically, at the same time that the demands for lower power envelopes using smaller and smaller batteries (hint: wearables) and longer times between charges are also getting more and more insistent. QuickLogic’s leadership recognized that their technology’s unique strengths may be perfect for addressing these conflicting demands.  

There is another set of conflicting demands creating a dilemma for OEMs, which QUIK’s capabilities can help solve. The companies that design and sell mobile devices face vicious competition in the mobile consumer device landscape (whether in smartphones, tablets, or the new and as-yet unproven market for wearable devices). They want to produce a device that has unique or differentiating features, with delightful built-in applications or algorithms that will make their device stand out from the others or make the owners of those devices happy and loyal. But they also have to bring those products out in a very short product-cycle, with “new generations” of phones and tablets being released about twice a year at the big top-five device companies (even faster than the cycle for new cars, for instance). Additionally, the OEM needs to be able to deliver these new, differentiating features in a cost-effective manner. So, they are facing a need for differentiation, inside this incredibly rapid product-cycle, with the need to be as cost-effective as possible so that they can offer the product at a competitive price.  These are very real and pressing concerns for OEMs, and once again QUIK’s fairly unique hardware background enables the company to offer some solutions that help solve this second OEM dilemma.

That’s two dilemmas OEMs are facing.

At the risk of oversimplifying somewhat, the aspects of QUIK’s silicon solutions that make them attractive to OEMs (because they help overcome the two dilemmas) include:

  • QUIK solutions offer programmable logic. Some of it is one-time programmable, and recently the company developed fully reprogrammable logic as well. This makes their solutions customizable and quick for an OEM customer to bring to market – even within the short product-cycles that have come to characterize the competitive, high-profile world of consumer mobile devices such as smartphones, and it also makes them cost-effective. In contrast, non-programmable preconfigured architectures such as ASICS would be prohibitively expensive to customize for most consumer devices, and the time it takes to design and bring them to market would also be prohibitively lengthy.
  • QUIK solutions also offer hard-wired “proven system blocks” or PSBs that they put right onto the same chip with the programmable logic. These PSBs are already hardwired and are not programmable: they resemble some of the characteristics of ASICs in that regard. These PSBs are hard-wired to perform certain tasks that QUIK has found many OEM customers want, but OEMs can basically choose from a “menu” of PSBs to add on to their chips. Moving some functions to hardwired PSBs lowers power dramatically, and also speeds up time-to-market, while still allowing some customizability through the use of the “menu.” By offering both programmable logic and PSBs together, QUIK can offer the advantages of both and help overcome the dilemmas described above, and in particular this set-up enables rapid time-to-market plus differentiating customizability plus cost-effectiveness, all at lower power than competing solutions.
  • QUIK offers processors which can offload specialty tasks from the central apps processor in order to save power, which will become more important as OEMs design more devices to be “always aware.” Running an apps processor constantly in order to keep a device “always aware” of its contextual surroundings would rapidly deplete the battery, because even as apps processors become more power efficient, they must by nature be flexible and capable of handling a wide variety of general purpose processing tasks, and fast enough to do those tasks in ways that will not annoy the user or delay the applications they are designed to handle, and therefore they cannot be optimized for power to the same degree as the more specialized processors such as those QUIK offers. The increasing demand that devices become more aware of their surroundings, which they do through the monitoring of sensor data, has already forced some of the leading-edge smartphone makers to incorporate a separate processor usually referred to as a sensor hub to monitor the sensor data and do certain types of processing to that data rather than making the apps processor handle it all. The sensor hub can monitor the sensor data and keep a low level of “always-on” awareness, and let the apps processor stay asleep until absolutely necessary, thus adding greatly to battery life and power efficiency. Apple introduced a sensor hub in the iPhone 5, for example.
  • QUIK offers a power-efficiency profile that is vastly superior to most competing solutions – orders of magnitude better in some cases, particularly as demand for processing becomes more significant (the magnitude of the improvement that QUIK can offer in this department is graphically illustrated by some graphs included in the more product-specific discussion below). One of the ways that QUIK’s solutions achieve much better power efficiency than competing architectures is their use of a different architecture which they believe offers much lower power-consumption. In fact, at this time there is data which suggests that all other architectures are facing a power wall that will make it difficult to continue adding sensor-enabled features, at the same time that forward-leaning mobile device OEMs are beginning to incorporate more and more sensors into their products. This paper from QUIK discusses the different architectures used by different processor designers on the market today, and why QUIK’s architecture offers greatly increased power-efficiency compared to other approaches.    
  • QUIK offers a robust set of software tools to enable OEMs to use QUIK’s hardware, further enabling them to use QUIK’s silicon and have it “talk to” and interface with other players inside the device. Some of these are “drivers,” which are analogous to the drivers we all use from time to time when we buy a new printer and want to enable it to interface with our laptop or desktop or tablet’s operating system. It wouldn’t be a very good idea for printer-makers to try to sell a piece of hardware that couldn’t interface with the software on a wide variety of other systems that their customers need it to be able to talk to. QUIK’s devices are programmable using C and C-like code which means that they can be manipulated, reprogrammed, tested, and debugged by many coders at the kinds of companies that would be using QUIK solutions, and that firms do not have to hire specialty personnel with skills in rarely-used coding languages in order to use QUIK products. Additionally, it also means that OEMs who use QUIK products will be able to transfer or “port” the code that they wrote for other QUIK products or for previous generations of QUIK products onto newer generations of QUIK products going forward into the future. There is a QUIK “roadmap” for customers to see the capabilities that are planned for future generations of processor solutions, so that customers can be confident that as they themselves want to introduce new generations of products, QUIK will have new solutions that will support new improvements and added functionality.
  • QUIK actually sends their own engineering teams to work together with the OEM’s engineers in order to make the integration of their products more successful and easier for the customer, especially for customers who need to design specific customized features. As has already been discussed, there is already a “menu” of “pre-loaded” features that customers can select in order to build a kind of unique and differentiated set of features, all of which are already basically “ready-made” by QUIK and don’t require a lot of design and added coding by the customer, but QUIK provides a full sales-support cycle to help customers ensure they get the solution that they are looking for.
  • In addition to the types of drivers and functionality-interface coding described above that QUIK has been offering on products for years, QUIK has now developed a specific set of sensor algorithms called the SenseMe library of algorithms, for use with their new sensor processing solutions targeting the nascent sensor processing market. This algorithm capacity is a major differentiator for QUIK in comparison to competitors who are offering a simple hardware sensor hub.  With a simple sensor hub, the task of either creating algorithms for the sensor hub to use in order to make use of the sensor data, or buying third-party algorithms from a software design company and putting those onto the sensor hub falls to the OEM company, some of which have their own algorithm teams (mostly only the top-tier OEMs) and many of which do not. By developing their own high-quality in-house algorithms, QUIK is offering much more than just a sensor hub product (even though their product is significantly lower power, faster time-to-market, etc). By virtue of their algorithm capability and sales-support tradition, QUIK is offering an entire solution and not just a hardware product – QUIK calls them “sensor processing solutions” as opposed to simply “sensor hubs.” It is a hardware and software solution, with design-integration assistance throughout the life-cycle of the sales and design process with the OEM customer. This adds yet another layer to what is already a very compelling value proposition. We believe QUIK’s products would be compelling enough just on the advantages of programmability and power-efficiency, but when the value of the algorithms and the “complete solution” approach is considered, we believe that QUIK could and will find some way to benefit from the tremendous added value that they are now bringing to the marketplace.

QUIK’s lineup of sensor processing solutions:
QUIK has several lines of programmable processors suited for the different general applications described briefly above, some of which allow the customer to write their own code and add their own unique functionality, and some of which offer pre-loaded PSBs so that customers can get “off-the-shelf” functionality for certain frequently-requested processing tasks. Interested researchers can look at discussions of these product lines on the QUIK website and in their various filings, whitepapers, and other publications. However, in this particular discussion we will focus primarily on QUIK’s most-important future growth opportunity, in our view: the sensor processing solutions designated by the letter “S” in the QUIK product nomenclature, and specifically their current-generation solution dubbed the ArcticLink-3 S2 (which we will simply refer to as the “S2” in this report), and the next-generation solution due out later this year, the ArcticLink-4 S3 (which we will likewise simply refer to as the “S3”).

As discussed briefly above, these processor solutions are designed to help OEMs who are designing mobile products that contain various sensors which need higher levels of monitoring and “data-fusion” tasks, particularly OEMs designing smartphones and wearables but also potentially tablets or other mobile devices containing sensors.  We are all familiar with mobile devices containing sensors: most smartphones and tablets, for instance, will “rotate the screen” based on their awareness of the way you are actually holding the device from one moment to the next, and the way they “know” is generally through data provided by a now-ubiquitous type of sensor called an accelerometer. Note, however, that this accelerometer does not need to be sending data to the device’s central processor (called the “apps processor” in the mobile world, or simply the AP) every second of every day: it only needs to do so when you are actually looking at the screen and the screen is lit up. For much of the day, whether the device is sitting on your desk or in your pocket, purse, briefcase or passenger’s seat, the screen may be off and the accelerometer and the apps processor can be resting.  

Because of this fact, device designers have often been able to get away with using no dedicated “sensor hub” at all – perhaps the apps processor could handle the sensor data when data is actually coming from the sensor, without draining the battery too rapidly or causing too much of a strain on the apps processor and its ability to handle other tasks. However, as devices begin to incorporate more and more sensors, and as OEMs and apps developers begin offering functionalities that people might find useful but which require sensors to be monitored more frequently or even monitored on some level all the time, it will become more difficult to use the AP alone for all this sensor data. This is because the AP for various very good reasons is usually not the most power-efficient processor type for these tasks, since it is optimized to run different kinds of apps for the device and the user.  It is for these reasons that some devices have already begun using a sensor hub. We will see that the kinds of sensor hubs in use today, while capable of handling the current sensor-related functions of this generation of smartphones, will probably choke on the levels of data that will be coming soon (or at least break the “power budget” of the battery, leading to bad product reviews from users who complain that the battery does not last as long as advertised and that they have to charge it up more frequently than they would like to, etc).

The QUIK S2 solution is designed to handle the types of sensor-data processing demands found on even the most advanced smartphones and wearables on the market today, and even to handle far more sensor data than what is available on any smartphone or wearable today (yes, even that one), and to do so with far better power efficiency than the commonly-used microprocessors which competitors (such as ATML, with their RISC-based M-series sensor hub microcontrollers, including the lower-end M0 and the state-of-the-art M4F) are offering. It can handle up to twelve sensors on a single device, and is capable of handling data sent from any of the most-common types of sensors available to OEMs today, including accelerometers, gyroscopes, thermometers, magnetometers, heart-rate monitors, pressure sensors, UV spectrum sensors, gas detecting sensors, humidity sensors, ambient light monitors, proximity sensors, and others.  It is fully compatible with most versions of the Android OS up to and including Lollipop 5.0, and with RTOS-based systems (“real-time OS”-based systems) 

The S2 uses only 93 microAmps (μA) at 1.2 volts. Remember that watts are the measure of power, and in this case are found by multiplying the amps of electric current times the volts of electric potential, meaning that at 1.2 volts the S2 uses 111.6 microwatts. Recently, QUIK announced a newer version of the S2 which uses even lower power, dubbed the S2 LP, which uses only 76 microAmps at 1.2 volts.

The amount of microwatts that the S2 will use inside an actual device will depend on the number of “instructions per second” that the processor has to process, which is a measurement of the load on the processor referred to as the “duty cycle.” The duty cycle required will be dependent on the number of times the sensor “samples” per second (measured in Hertz, which is simply the number of times something happens in a second) and the algorithms that the processor is running, and more efficient algorithms might be able to perform a task using fewer instructions per second than a less-efficient algorithm (QUIK believes its algorithms are proving to be some of the more efficient algorithms for many sensor tasks). But, for an algorithm that requires one million instructions per second, or “1 MIPS,” the QUIK S2 LP sensor processor would use a little over 150 microwatts, while the M4 would use over 200 microwatts. The chart shown below compares the power consumption of the QUIK S2 LP to the ATML M4F for duty cycles up to 1 MIPS (these charts are available on the company’s IR page under “financial events” from the earnings call and presentation for the quarter ending March 2015):



While the S2 LP has a lower power-consumption profile, the M4F can basically handle these tasks within the power envelope of most smartphone batteries today without breaking the power budge of 1% to 2% of the total battery charge that OEMs have said they can tolerate being depleted by a dedicated sensor processor. This fact explains why OEMs today can use microcontrollers such as the ATML M4F and satisfy the sensor data loads they need to process within their existing power budgets.

However, more demanding “always aware” applications in future devices with more sensors sampling more frequently and running more algorithms (such as those monitoring your sleep patterns all night, or performing indoor navigation functions) will place higher demands on the sensor processing platforms, and QUIK believes that this is where the QUIK advantages will really begin to show versus competing approaches. The chart below picks up where the above chart leaves off – and looks at duty loads above 1 MIPS:



This chart projects the power consumption when using the QUIK next-generation sensor processing solution platform, dubbed the S3 and due for release later this year (“mid-2015,” for what that is worth). Clearly, if these projections are accurate, the differences in power uses will be so significant for more-demanding sensor-processing tasks that the S3 will be the clear choice for designers.  At 55 MIPS in the chart above, the S3 is using only 1.65 milliwatts, while the M4F is using somewhere above 10 milliwatts. Put another way, at 55 MIPS, the S3 can operate at a level of power consumption that the M4F hits at only 8 MIPS.

Again, it bears repeating that QUIK’s solution provides not only significant hardware-based advantages as a result of their unique architectural approach, but also a range of very efficient sensor-oriented algorithms, including some of those listed in the blue squares along the lower edge of the chart above, as part of their “SenseMe” library of algorithms.

Current customers:
Although it should be emphasized that the entire market for the kinds of sensor-rich, context-aware applications that these solutions are designed to enable for OEMs is still very much in its infancy, there are reasons to believe based on what QUIK has revealed so far that customers are extremely interested in the capabilities described above and that they are actively engaged in the evaluation stage on QUIK’s S2 product, with some of them lining up to evaluate the S3 as soon as it begins to roll off the fabs as well.

QUIK has three customers announced in the conference call from the December 2014 quarter who are using the S2 in a sensor-hub capacity in wearable products: one of them an eye-mounted device (from a company in Japan called Telepathy, on a product called the Jumper), and two of them wrist-mounted (one of them from FoxConn, on a product which we believe will be marketed in China, possibly for children to wear to enable them to communicate with their parents or enable parents to see their location, and one of them from a yet-to-be-named OEM). In the most-recent call for the March 2015 quarter, QUIK revealed that these companies all delayed their production for reasons unrelated to the sensor processing solution, although it appears that FoxConn did express interest in a new algorithm that QUIK is developing to add to their SenseMe library. Whatever the ultimate level of success these initial OEM customers have with these particular products, we believe these are likely just the very first little lapping waves on the shore of a trend in sensor processing products and requirements that will soon become a flood tide. Additionally, QUIK announced in the earnings call for the March 2015 quarter that these three customers have been joined by two more who are using the S2 in a sensor processing capacity and who are taking production-volume shipments in the current quarter (the quarter ending at the end of June 2015).

At the same time that these initial S2 customers are beginning to take production-level shipments from QUIK, other potential customers are engaged in various levels of evaluation of the S2. At the end of last quarter (ended December 2014), the company announced that engagements had jumped from around eight to almost two dozen, and during the most-recent quarter’s earnings call (for the quarter ended March 2015) the company revealed that this two dozen number has now nearly doubled, implying a very robust “pipeline” of potential future designs that could incorporate the S2 (close to fifty, in other words, and probably above 45 engagements).

Additionally, the company is offering an “alpha” program for its S3, which is something we have not seen from QUIK before on previous products. In this program, a potential customer who wants to evaluate the S3 would be sent versions (“rough drafts,” so to speak, for them to “proofread”) of the new product almost as soon as the QUIK engineers receive them from the fabs. They would receive these early runs with the understanding that some bugs still need to be worked out, but they would have the opportunity to use these QUIK products as part of their own design and evaluation on the way to possible incorporation in a smartphone or wearable or other mobile product. In order to be in the alpha program, a potential customer must have executive sponsorship (indication of interest in the S3 from the executive management, the customer must be a “top-tier device OEM” (such as LG, or Lenovo/Moto, or HTC, or XiaoMei, for example), they must have a specific funded project for which the S3 is being evaluated, and they must agree to a formal feedback process with QUIK as they receive and evaluate these S3 runs (QUIK engineers, of course, would be checking out the early products at the same time as they roll out of the fab). QUIK stated that they have one top-tier OEM in the program now, with a second in the engagement process as another potential participant in the alpha program.

Obviously, we do not anticipate S3 revenues from production-volume shipments to begin until some months after the first S3 products are released later this year – and thus probably not until calendar-year 2016. However, despite the company’s conservative guidance for the current quarter (at a guide of only $5.5 million in revenues, less than the quarter they just reported, this projection does not incorporate major volumes of S2 shipments this quarter either), we do believe that S2 shipment volumes will increase later this year, although the timing is dependent upon OEM customer business decisions and of course on the success of their various products.

We believe that the bigger point is that the market for wearables, and for all kinds of mobile devices including smartphones, containing sensors with more robust and more context-aware higher-MIPS algorithms is only in its very early stages, and that it is very likely that processing demands will increase dramatically in future months and certainly in future years. While the company only has five products shipping early stages of production volumes of the S2 this quarter, it is not unlikely that we could see ten to twenty products within just a quarter or two.  When (and if) a major smartphone OEM incorporates the first S2 or S3 in a flagship smartphone, we could see tens of millions of such devices shipping in a fairly short period of time (perhaps by a year or two from today). 

The company ended the most-recently-reported quarter with $28.2 million in cash, and $34.3 million in current assets.  Even at the low quarterly revenues of around $6 million, the quarter’s net loss was about $3.6 million – which means that at around $11 million in revenues the company would start to be profitable (perhaps a bit higher, as expenses would of course increase to meet higher volumes of production and to field higher numbers of sales and support personnel).  At this time we do not believe an additional capital raise to “keep the lights on” is likely to be necessary in the quarters between now and the time that revenue increases should materialize.

However, we realize that the Street right now does not care very much about all of the charts and predictions just described: they care about revenues and about one quarter of guidance. We believe this creates an opportunity for investors with longer time horizons (we generally measure our investment horizon for growth companies in terms of years rather than months or even quarters, especially for smaller growth companies such as QuickLogic).

Potential concerns:
Of course, as Benjamin Franklin famously admonished, “there’s many a slip, ‘twixt the cup and the lip,” meaning that all of these potentially lucrative markets and projections are still “in the cup” and there are potential pitfalls before they can be realized.

We will address just a few of the most important concerns investors must consider.

  • Alternative solutions: the discussion above shows the significant potential advantages QUIK is bringing to the market. However, technology investors know that creative people can come up with other ways of solving problems, and there have been instances where the “best” technology did not in fact win out in the marketplace. Some argue that apps processors will be able to absorb the role that dedicated sensor data processors are filling today, although we believe there are several reasons that the direction AP designs are going will make them even less likely to be able to handle sensor processing and “always aware” functionality than they are today. 
  • Smart sensors: of greater concern perhaps is the idea of the “smart sensor,” or the incorporation of “sensor hub” functionality into a ubiquitous sensor such as the accelerometer (accelerometers are now present on virtually every smartphone and will probably be found on most wearables). Why would OEMs want to buy a separate sensor hub if they can get similar functionality from a “smart sensor”? This is a legitimate question, and we believe some OEMs may go this route for some devices. The fact is, however, that the potential market is likely to be very big, and very diverse, and while this solution may be appropriate for some devices, that does not mean that it will prevail in all cases, or even in most cases. We believe that it may not be attractive to most OEMs because there are several sensor companies out there, and OEMs like to have them competing for slots in the OEM’s products, so that the OEM can have multiple suppliers competing against one another and multiple suppliers from which to choose (rather than getting locked-in by one sensor provider).  Sensors may be somewhat commoditized, and sensor companies would like to get out of that category, and so of course they like to promote the “smart sensor” approach, but going this route could lock the OEM in to one sensor provider not just for current products but – as we have touched upon – for their “future roadmap.” This is a serious counter-argument to the “smart sensor” approach, in our view. Again, we believe it may be used for some portion of the market, but that the QUIK solution is not unduly threatened by this possibility.
  • Other programmable companies: Of the other makers of programmable logic, Lattice (LSCC) has been the most focused on addressing the needs of the mobile market, but they have focused more on the types of “workaround” applications that programmable processors have historically been good at solving for “shorter-term” product designs. At this time, we believe the QUIK distinctions in terms of their PSB approach, their specific architectural decisions to maximize power efficiency even at higher duty cycles, and their “complete solution” approach of developing a library of in-house algorithms provide QUIK with clear differentiation at this time. 
  • Major unforeseen battery improvements, or battery-charging improvements, such as the ability to simply charge the battery all the time wirelessly wherever you are.  So far this is not a serious concern, but it could become a concern in the future.
  • An inability by the company to handle large orders when they do begin to arrive. We believe the team is experienced, solid, and competent. They have been through the vetting process at Samsung and have had Samsung as a customer for years now. We believe that large orders will be a “good problem to have” and that QUIK will be able to handle those orders with their fabless model when those orders begin arriving. Nevertheless, this is always a good concern but one of which investors should be aware. A failure to ship product to a major OEM can ruin a company’s reputation very rapidly, and we have seen it happen to small companies that take years to recover or which in some cases never really recover at all.
  • A major reversal in the use of sensors, or a rejection of sensor functionality by the public. 
  • A failure for the increase in sensor data sufficient to demand duty cycle loads requiring more power-efficient sensor data processing. The QUIK sensor processing solutions become more attractive at higher duty cycles, and at this time may be the only possible solution for duty cycles above some threshold between 25 MIPS and 50 MIPS. However, if data processing requirements never reach those thresholds, the advantages of QUIK’s power efficiency are still quite significant, although less dramatic than at higher duty cycle levels. If sensor data loads never approach those levels, it would impact the investment thesis. However, we believe this possibility is very unlikely. The reader can conduct his or her own research on the potential market for wearables, and can consider the types of competition already present in the smartphone market and the desire for OEMs to be able to offer features that differentiate their products and help their phones impress potential buyers.  Some have noted that the “internet of things” can in many ways be thought of as an “internet of sensors,” so central will sensors be to the types of tasks that wearables and other low-power, mobile and connected devices being dreamed about, designed, and developed right now. And, while we believe the Street has yet to truly realize it in any meaningful way, we believe that virtual reality or VR has the potential to be truly transformative for a wide variety of visual media, education, entertainment, gaming, and sports broadcasting. VR devices such as the Samsung VR will require additional sensors and additional sensor data processing. In short, while the forecasts for the types of new devices and new requirements that are discussed above are still very tentative, it is very likely that this market will grow exponentially over the next few years.
  • Arguments that the management team is not competent. Some bearish pieces, at least one of them from a short-seller of the stock,  have made this argument. We believe that not only is it incorrect, it is backwards. In our assessment, CEO Andy Pease arrived at QUIK in 2006, became CEO in 2011, and made strategic decisions regarding the company’s specific strengths, and regarding the  new markets where QUIK’s specific strengths could offer the greatest value. The decision to target the incipient sensor data processing needs of mobile OEMs and engineers we believe to be frankly a brilliant application of QUIK’s unique strengths to an area where OEMs and engineers are facing two very real and very pressing dilemmas, both of which QUIK is uniquely positioned to solve. If the company had to use shorter-duration “connectivity” and “display-bridging” revenues as “stepping stones” to get them to the point where they are today, positioned in front of a nascent sensor data processing market that is just now in the beginning stages of opening up, then we believe that was an example of good management and good leadership.
  • The resignation of their CFO. We are sorry to see their CFO depart for a new opportunity, but in our opinion the CFO does not set the vision for the company. The CEO sets the vision and the CFO helps make it happen. We do not see any indication that the CFO’s departure had anything to do with an inability to make it happen, but rather that it had to do with an immediate offer from another company which he decided to accept for his own reasons that have to do with what is best for him. We do not believe that it will be difficult for QUIK to find a competent person to manage the company’s finances in a way that will support the current vision. We believe the current vision for the company is clear, that it is a valid strategy, and that the financial conditions in terms of cash flows are probably going to be sufficient to achieve it. The personality of one CFO versus another will not change that at this particular juncture, in our opinion.

Conclusions:
We believe that the QUIK story is fairly complex and requires a rather involved level of market-specific knowledge, as well as familiarity with some specialized semiconductor architecture issues. The company is currently not widely or well understood, in our opinion. Investors have recently been seriously spooked by declines in revenues from products that were all along necessarily shorter-duration designs with OEMs, in bridging and connectivity applications. The real growth opportunity for this company, we believe, is in the sensor data processing market, which is still so early that it is impossible to predict with great detail, but which will almost certainly grow exponentially over the next two to seven years. Investors who understand this, however, have recently also been disappointed by the failure of the sensor market to immediately ramp, and have punished the stock accordingly. The ride in this stock over the past year has been extremely painful, as highs that were generated based on excitement over shorter-duration products evaporated rapidly and the stock has seen several sharp freefalls to successively lower lows since then. 

However, we believe that the sensor opportunity is still just getting underway, and that the company’s customer engagement numbers reveal a robust pipeline. In fact, the company is in the enviable position of sitting in front of a potentially tremendous market in consumer mobile devices (products with some of the highest potential volumes on the planet), at a time when it could be facing one of those fabled “paradigm shifts” which one reads about in business school or related literature, but which investors at times stop believing in. With a market valuation of just under $100 million including cash, and with cash on hand (at end of March 2015) of just over $28 million, QUIK is a company that is basically valued around $70 million – which may indeed be an incredible value, based on the potential markets just described and the potential advantages their solutions offer to OEMs. 

In many ways they are like a venture investment, which often price at valuations this low in their early investment rounds, on companies that still don’t even have revenues (QUIK is set to get at least $20 million a year) or breakeven potential for years into the future (QUIK might reach breakeven in a matter of a few quarters) or major customers (QUIK has had Samsung as a customer, and appears to have one if not two other top-tier OEMs considering their upcoming S3 product for evaluation). Like a venture investment, there are risks, but also like a venture investment, there are eye-opening possibilities to consider, especially at these valuations. And, unlike a venture investment, QUIK is a public company, filing financials with the SEC in accordance with GAAP, and trading on public exchanges where there is immediate liquidity for investors if they need it (unlike a venture investment which can require lock-in for years). Unlike a venture investment, QUIK is also available to any investor who wants to buy it – although of course we recommend using this report as only a starting point for your own evaluation and due diligence of the investment merits of the company. 

And finally, unlike many venture investments, QUIK is led by a seasoned management team with a board of experienced professionals who know the company, know each other, and know the markets (the board includes QuickLogic’s founder and other industry veterans).

Important Disclosure Information: 
At the time this report was submitted for publication, the principals and clients of Taylor Frigon Capital Management owned securities issued by QuickLogic (QUIK). Of all the other companies mentioned in the report, at the time of submission for publication, the principals and clients of Taylor Frigon Capital Management owned securities issued by Apple (AAPL) only, and none of the others.


Please remember that past performance may not be indicative of future results.  Different types of investments involve varying degrees of risk, and there can be no assurance that the future performance of any specific investment, investment strategy, or product (including the investments and/or investment strategies recommended or undertaken by Taylor Frigon Capital Management LLC), or any non-investment related content, made reference to directly or indirectly in this research will be profitable, equal any corresponding indicated historical performance level(s), be suitable for your portfolio or individual situation, or prove successful.  Due to various factors, including changing market conditions and/or applicable laws, the content may no longer be reflective of current opinions or positions.  Moreover, you should not assume that any discussion or information contained in this research serves as the receipt of, or as a substitute for, personalized investment advice from Taylor Frigon Capital Management LLC.  To the extent that a reader has any questions regarding the applicability of any specific issue discussed above to his/her individual situation, he/she is encouraged to consult with the professional advisor of his/her choosing. Taylor Frigon Capital Management LLC is neither a law firm nor a certified public accounting firm and no portion of the research content should be construed as legal or accounting advice. A copy of the Taylor Frigon Capital Management LLC’s current written disclosure statement discussing our advisory services and fees is available for review upon request.
Continue Reading