Good day, ladies and gentlemen, and welcome to QuickLogic's Fourth Quarter and Full Year Fiscal 2018 Conference Call. At this time, As a reminder, this conference is being recorded. Would now like to turn the conference over to your host, Ms. Mariah Shilton, with Investors Relations. Ma'am, you may begin.
Thank you, Valerie. Welcome, everyone, and thank you for joining us today. The QuickLogic's 4th quarter and full year fiscal 2018 results conference call. With us today are Brian Faith, President and Chief Executive Officer and Doctor. Sue Chang, Chief Financial Officer.
Before we begin, I will read a short safe harbor statement. Some of the comments QuickLogic makes today are forward looking statements that involve risks and uncertainties, including, but not limited to, stated expectations relating to revenue from new and mature products, statements pertaining to QuickLogic's future stock performance, design activity, then its ability to convert new design opportunities into production shipments, timing and market acceptance of its customers' products. Schedule changes and projected production start dates that could impact the timing of shipments, the company's future evaluation systems, broadening the company's ecosystem partners, expected results and financial expectations for revenue, gross margin, operating expenses, profitability and cash. These statements should be considered in conjunction with the cautionary warnings that appear in QuickLogic's SEC filings. For additional information, please refer to the company's SEC filings posted on its website and the SEC's website.
Investors are cautioned that all forward looking statements in this call involve risks and uncertainties and that future events may differ materially from the statements made. For more details of the risks, uncertainties and assumptions, please refer to those discussed under the heading Risk Factors, in the annual report on Form Ten K for the fiscal year ended December 31, 2017, the company filed with the SEC on March 9, 2018. These forward looking statements are made as of today, the day of the conference call. And management undertakes no obligation to revise or publicly release any revisions of the 4 looking statements in light of any new information or future events. Please note, QuickLogic uses its website, the company blog, QuickLogic hotspots, its corporate Twitter account, Facebook page and LinkedIn page, as channels of distribution of information about its products, its planned financial and other announcements, its attendance at upcoming investor and industry conferences and other matters.
Such information may be deemed material information and QuickLogic may use these channels to comply with its disclosure obligations under Regulation FD. The conference call is open to all and is being webcast live. We will start today's call with the company's strategic update from QuickLogic's CEO, Brian Faith. Then CFO, Sue Chung, will provide financial results and guidance. Brian will deliver closing remarks and open the call to questions.
At this time, is my pleasure to turn the call over to Brian Faith, President and CEO. Please go ahead, Brian.
Thank you, Brian. Thank you all for joining our Q4 2018 conference call. This is a very exciting time for QuickLogic. We are building traction in our core markets, and layering on to that with our more rapid penetration in AI. The enthusiastic reaction from our partners, customers and distributors following our acquisition of SensiML has been very encouraging.
They acknowledge that there is pent up demand for an end to end solution that we now provide Large customers are beginning to move EOS S3 designs into production and other key designs are scheduled to ramp during the first half of twenty nineteen. The Master Technology license agreement or MTLA go to market strategy for our embedded FPGA IP continues to build traction and the customer and partner reception for our QuickAI and SensiML strategies have exceeded my expectations. I'll get into all this in much greater detail, but first let's take a couple of minutes to review 2018. There were 3 factors that significantly reduced our new product revenue for Q4. First, the decline in Q4 display bridge and connectivity revenue was much larger than we had originally modeled.
That will weigh on total new product revenue. 2nd, a government contractor that we have been working closely with an embedded FPGA IP, MTLA and license agreement did not receive its anticipated government funding and time to execute the agreements in Q4. 3rd, while we recorded our first material QuickAI revenue in Q4, one of our QuickAI customers made a last minute design change to our new EOS S3AI solution from the originally anticipated module solution. This lowered the ASP and resulting revenue from the sale. In our February 2018 conference call, I forecasted our baseline revenue for 2018 would be approximately $10,000,000 In that call, I define baseline revenue as revenue from mature products, display bridges and existing connectivity designs.
For full year 2018, our total baseline revenue was in line with our forecast. However, within the baseline, higher than expected mature product revenue offset a significantly larger decrease in non strategic new product revenue than we anticipated. This resulted in a decline of total new product revenue for 2018 This shift towards strategic business is reflected in our full year non GAAP gross profit margin of 51%. Which is up significantly year. While this analysis more accurately portrays the progress we've made towards realizing our 2018 goals, than the aggregate decline in new product revenue shown on today's press release.
The growth in strategic business was below the expectations we shared a year ago. Looking forward to 2019, we anticipate mature product revenue will be flat to up slightly with 2018, and that non strategic new product revenue will decline by approximately $1,500,000. With that in mind, The question I'm sure all our investors have is what has changed? My goal today is not only to answer that question, but also share with you some exciting new developments that I think will help you better understand the value of our And with that, share my belief that we are very well positioned to deliver greater than 50% revenue growth in 2019 and continue our established trend embedded FPGA IP business. After determining that our original go to market strategy resulted in us getting stuck in a couch 22 loop, We changed our strategy in mid-twenty 18 by offering our targeted customers the interim step of an MTLA.
Since then, we have signed 3 MTLA's and have a 4th that is pending government funding. The MTLA strategy has benefited us and our potential customers in many ways. It created a low cost method for our targeted customers to develop test chips that they could use to evaluate the performance and benefits of embedded FPGA. It outlines all the criteria for subsequent embedded FPGA license agreements. This allows design groups to quickly and easily license our embedded FPGA IP for a targeted SoC or ASIC without the need or associated delays of extensive legal and executive level reviews.
It also enables us to much more quickly and efficiently qualify whether or not a given engagement is likely to result in a license agreement. C Sky, which was subsequently acquired by Alibaba. We believe this MTLA, which carries a modest quarterly maintenance charge will lead to multiple SoC license agreements beginning this year. We signed our second MTLA with ETH Zurich. Which was also licensed to develop its new RISC V Parallel Ultra Low Power or PULP Platform, using GlobalFoundries Advanced 22 FDX fabrication process.
Takeout of the Pulp platform incorporating our embedded FPGA IP was completed during Q4 2018, and we expect to receive 1st silicon in Q2 2019. We signed our 3rd MTLA with a semiconductor company and recognized several $100,000 in revenue during Q4. We expect this will be the first of many SoC license agreements with this company. While our NDA with this customer prevents me from adding more color, I'm confident that once the details become available, you'll appreciate the significance and value of this agreement. When we discussed our outlook for Q4 in our last quarterly conference call, we expected to sign a new MTLA and recognize revenue in Q4 from an ASIC license agreement with a government contractor.
However, due to delays in government funding, the agreements were delayed. We've worked closely with this customer for a long time, and we are confident this agreement will be executed. As it stands today, We anticipate the MTLA and license agreement will be finalized during the first half of twenty nineteen. During Q4, we also expanded qualifications with our fabrication partners by adding TSMC's 40 nanometer node. With this, We now have our EFPGA available at TSMC for 65 nanometer and 40 nanometer at global foundries for 65 nanometer, 40 nanometer ended unique 22 nanometer FDSOI nodes and at SMIC for its 40 nanometer fabrication node.
There's a good chance we will expand to additional process notes during 2019, but that expansion is most likely to be driven by customer SoC or ASIC license agreements. We will begin rolling out several new initiatives this quarter that we believe will not only add to the momentum we have established with our MTLA go to market strategy, but also improve the scalability of our engagement process shorten the time it takes us to qualify the likelihood of winning an engagement and provide cross leverage for our other IP and product offerings. The ETH Pulp platform will enable us to quantify the benefits of running offload engines like accelerators in our embedded PGA versus running them in software on the risk 5 processor. The SensiML Analytics Toolkit enables designers to run AI models that offload neurons to our embedded FPGA. Here, we can leverage the SensiML Analytics Toolkit in conjunction with EOS S3, to demonstrate how embedded FPGA can improve system performance and free up processor resources.
This can be used to add new features or left idle to lower power consumption. During Q1, we will release a new application programming interface or API that will enable software engineers to leverage the inherent benefits of embedded FPGA. The API is written in an industry standard format and enable software engineers to write code as they normally would and then easily evaluate the performance and or power savings benefits of running math intensive operations in embedded FPGA. In conjunction with the API release, we will also introduce our hardware accelerator library. This will include tested software code to instantiate EFPGA based accelerators optimized for voice recognition applications.
We will follow this up with additional accelerators that are commonly used in time series AI applications supported by the SensiML Analytics Toolkit. With these new tools and initiatives, we will Sensible SaaS license and EOS platforms. Improve the scalability of our engagement process which will increase the number of engagements to understand and quantify the value of embedded FPGA more easily, shorten the time it takes us to qualify the potential of an engagement to produce an MTLA and revenue generating licenses, and I believe increase the percentage of engagements we win. Let's turn now to update the status ES 2018 and Amazon forecasting it would release its close talk specification for always onalways listening devices any day We were very optimistic these and other similar designs in our pipeline would contribute significantly to our 2018 revenue growth. However, the close talk specification wasn't publicly released until last October and the test certification files used in conjunction with the spec have been updated several times since.
The headwinds we faced last year weren't just from Amazon though. We move through 2018, we continue to win designs, including a major platform design with a large Japanese smartphone company, only to see production forecasts move forward often a quarter at a time. The short story is design engineering groups and even product managers were fairly quick to see the benefit of always onalways listening devices. However, with still solid demand for push to talk, companies were hesitant to move our next generation design wins into production. As a result, outside a few specialty applications, the vast majority of our design wins that sat in a log jam last year were ready, but waiting for a production schedule.
Underscoring the delays we faced last year, a design we won in early 2018 with JD dotcom Finally, moved into production just last month and was proclaimed by future source consulting as being the industry's first always on wake word Bluetooth in ear consumer headset. In addition to the design with JD dotcom, 2 of the designs we displayed at CES 2018 have moved into mass production. Clear has initiated production its Amazon AVS compliant battery powered stage speaker. And one of our ODM headset designs, which originally targeted Amazon, won a significant contract with SF Express. With the log jam clearly loosened and the Amazon close talk spec in hand, We anticipate a number of our other hearable design wins from 2018 to move into production during the first half of twenty nineteen.
I'm not going to tell you the logjam that stunted our growth last year has been totally cleared. We still don't have committed schedules for all of our design wins. However, The trend is definitely moving in the right direction and with several of our largest customers now committing to production schedules, the momentum seems to be building. Last year, a major Japanese smartphone company selected our EOS S3 for a platform design it intends to use in several new models. Following this, we negotiated a broad MOU with this company that covers all MCU applications and smartphones, feature phones, and IoT devices.
This is a big win for us. We originally expected the first smartphone to move into production in Q3 2018. However, the primary wireless carrier for the design asked our customer to push it out to 2019. This represented one of the several significant headwinds we faced 2018. Things are moving forward well with this customer and we anticipate shipping meaningful EOS S3 volume during Q1 to support initial production in a new always onalways listening application displayed the 1st model incorporating the design at CES last month.
The customer constructed an environment more similar to the There were two reasons the customer chose to do this versus demonstrate the feature on the CES floor. First, The schedule released to retail outlets was pushed out to July 2019. And second, the floor at CES is a very noisy environment. We currently plan to initiate production shipments for the We believe additional OEMs will introduce the platform design in various models during the second half of twenty nineteen. In addition to these design wins, we are also engaged with the original platform design customer on a second EOS S3 application.
The engagement with a large consumer goods company that I mentioned last quarter is progressing design that it will move into production during the first half of twenty nineteen. We are also working closely with this OEM as it evaluates EOS S3 for a new platform design that targets the variety of high volume consumer products, all of which would represent new product categories for QuickLogic. On January 17th, JD dot com released its new pilot U LIFE N1 smart headset with EOS S3 enabled always onalways listening capabilities. With over 300,000,000 active customers and over $60,000,000,000 in 20.18 revenue, jd.com is China's largest online electronics retailer. With the ultra low power consumption of EOS S3, jd.com was able to limit battery size and weight to only 30 grams and still deliver 90 hours of standby time on a single charge.
Even more impressive, according to Chinese Tech Newsource EK21, the pilot U Life N1 provides a voice recognition rate of 95%, which is far better than the recent analysis by 9 to 5 Mac the rate series accuracy at only 78%. This could explain why after only days in the market, There were thousands of reviews posted for the pilot with wet EK21 term as a praise rate over 98%. SF Express adopted 1 of the ODM hearable designs that we displayed at CES 2018 for its new Shunfeng smart Bluetooth headset. This all weather design enables hands free access to the company's database as well as cloud based AI resources to accelerate decision making. SF Express is one of the world's leading logistics companies and has said it intends to deploy the headsets to its 600,000 delivery agents in China.
The company estimates that the hands free voice interface enabled by EOS S3 will improve its efficiency by 70%. Let's shift now to QuickAI and SensiML where things continue to come together more quickly and efficiently than we originally modeled. As I noted earlier, s3AI SoC platform, which has a lower ASP. EOS S3AI can be used as an SoC running a soft neural network in the MCU and EFPGA, or for more compute intensive applications in conjunction with a neural net processor like the nepes and M500. Please see the press release that was issued after the markets closed today for more information on EOS S3AI.
One of the more frequent questions I have fielded from investors since our conference call last November has to do with our forecast to ship material QuickAI revenue nearly 1 year sooner than we had originally forecasted. The short story comes in two pieces. First, we underestimated the pent up demand for a market ready AI endpoint solution. There are obviously a lot of companies talking about AI but surprisingly few have market ready solutions that are practical to deploy today for Edge and Endpoint applications. The second reason is, together with the SensiML Analytics Toolkit, we offer customers not only a market ready solution, but even more importantly, an end to end solution that is quick and easy to adopt without the need for significant data science or even firmware engineering resources.
The EOS S3AI platform is a complete hardware and software solution that is ideal for endpoint applications such as predictive maintenance, structural health monitoring, manufacturing process control, and context awareness for wearables. This single chip solution leverages the SensiML Analytics Toolkit to efficiently identify which complex feed extraction algorithms are ideal for the customer labeled datasets and generates the AI models and classifiers using machine learning techniques for code development. The AI model that is subsequently generated is uniquely optimized and partitioned to fit in the EOS S3 AI hardware blocks. Which included embedded FPGA flexible fusion engine and MCU. Following this, the devices program seamlessly by the SensiML Analytics Toolkit.
This tight integration of hardware and software ensures that the AI algorithms created with by the Toolkit are optimized for the EOS S3AI platform, enabling low power and efficient endpoint processing. We are working closely with It has been a long time since QuickLogic has offered something that is attractive for our channel partners and distributors to market proactively. But with QuickAI and SensiML, that has changed and our partners are excited. To build on this early momentum, we have combined our QuickAI Hardware Development Kit or HDK, with SensiML's analytic toolkit to operate very efficient and low cost way for OEMs to evaluate the benefits of AI. And quickly integrate our unique solution into new designs.
The customer gets everything they need to develop and test their designs. If the customer likes the result, we end up with both a SensiML SaaS license and an new production design win for EOS S3AI. We have already rolled out this strategy and our new HDK in Japan, Taiwan, South Korea, the U. S. And some European countries.
In fact, our Japanese distributors have already committed to approximately 50 HDKs in this quarter. With the support of our distributor network, this is vastly scalable and a near 0 touch initiative for QuickLogic. In addition to our channel partner strategy, we are also leveraging This layers on very well with the engagements and customer commitments SensiML had already established prior to our acquisition. Some of which are with Fortune 500 companies. The short story here is we have not only hit the ground running, we have done so with good traction.
I would now like to turn the call over to Sue for discussions of the financials. Sue?
Good afternoon, and thanks to everyone for joining us today. Please note, we are reporting our non GAAP results You may refer to the press release we issued today for a detailed reconciliation of our GAAP to non GAAP results and other financial statements. We have also posted an updated financial table or IR webpage provides the current and the historical non GAAP data. For the fourth quarter of 2018, Total revenue was $3,200,000 within our forecasted guidance range. Our new product revenue was one point $3,000,000, reflecting a higher than anticipated decline of our display bridge and the connectivity revenue.
And a customer design change from a quick AI module solution to an EOS S3AI solution. Our mature product revenue was $1,900,000, reflecting higher demand from our military aerospace and defense customers. Our Q4 2018 gross margin was 52.6%, within our forecasted range. Excluding a one time inventory reserve or Q4 gross margin would have been 68.2%, reflecting the benefit of our EFPGA IP raises, QuickAI and higher than anticipated mature product revenue generated in the quarter. Operating expenses for Q4 or $4,300,000 below our forecasted range due to that later than anticipated timing of engineering new hires.
R and D expenses were $2,300,000 and SG and A expenses were $2,000,000. The total for other income expense and taxes in Q4 twenty eighteen was a $13,000 charge. Which was below our forecast due to capital gains from the sale of a foreign depreciated manufacturing equipment. Net loss was $2,600,000 or $0.03 per share. Net cash usage during the 4th quarter was $3,800,000 within our forecasted range.
Cash usage was negatively impacted by a net increase in working capital, driven mostly by endofthequarter shipment that increased the accounts receivable by nearly $1,000,000. During the quarter, we increased our line auto credit from $9,000,000 to $15,000,000 at the same interest rate. Our cash balance at end of the year was $26,500,000, which includes a $15,000,000 loan from the revolving line of credit. Turning to the full year 2018 results. Total revenue was $12,600,000, up 4% compared to $12,100,000 in fiscal 2017.
New product revenue for fiscal 2018 was $5,700,000, down 2% compared to 5.8 $6,900,000, up 9% compared to $6,300,000 in fiscal 2017. Due to our continued success in diversifying our customer base, we had multiple customers with a greater than 10% for the full year. Gross margin for 2018 was 51.2% up from 46.4% in 2017. The higher gross margin in 2018 was primarily driven by mature products. Operating expenses for fiscal 2018 were $18,200,000, which were flat with the prior fiscal year due to modest total revenue growth and our higher gross margin, our net loss for 2018 decreased to $11,900,000 or $0.14 per share compared to $12,700,000 or $0.17 per share for 2017.
Now let's turn to the first quarter 2019 forecast. Our revenue guidance for Q1 2019 is approximately $3,500,000, plus or minus 10%. Total revenue is expected to be comprised of approximately $1,500,000 of new product revenue and a $2,100,000 of mature product revenue. The increase in mature product revenue is due to continuing higher demand from our military, aerospace and defense customers. The increase in new product revenue is anticipated to be muted by further decreases in our display bridge and the connectivity revenue.
On a non GAAP basis, we expect our gross margin to be approximately 55%, plus or minus 3%. We're forecasting non GAAP operating expenses at approximately $5,100,000 plus or minus $300,000. We expect our non GAAP R and D expenses to be approximately $3,000,000, which includes The new employee is obtained through our SensiML acquisition. Non GAAP SG and A expenses will be approximately $2,100,000. We expect our other income expense and taxes will be a charge of approximately $60,000.
At the midpoint of our forecast, our non GAAP loss is expected to be approximately $3,200,000 or $0.03 per share. As was the case in prior quarters, the main difference between our GAAP to non GAAP results is our stock based compensation expense. Which we expect to be approximately $835,000 for the first quarter. Out of which $210,000 will be a one time charge related to our all stock acquisition of SensiML that closed in the beginning of the year. In Q1, we expect to use between $3,300,000 $3,800,000 in cash, reflecting the timing of working capital and increased R and D spending.
I'd like to note, in line with good corporate governance practices, we will be refreshing our $40,000,000 shelf registration in the coming weeks. We will file Form F3 right after we filed the 10 K. With that, let me now turn the call back over to Brian for his closing remarks.
Thank you, Sue. 2019 is setting up to be a very good year for QuickLogic. Our EOS S3 design wins are finally moving into production and new strategic design wins at large OEMs are scheduled to begin ramping in the first half of twenty nineteen. Our MTLA go to market strategy is working. And beginning this quarter, we are poised to introduce our open API strategy accelerator library and have the recognized nearly a year earlier than we originally expected.
QuickAI is already building serious traction and we are taking steps to further leverage what we believe is a very unique solution. SensiML not only puts us in the very solid position of being in control of our own destiny, but also enables us to offer customers the convenience and security of one stop shopping for a practical end to end solution for AI in Edge endpoint applications. With the QuickAI platform and SensiML Analytics Toolkit, we have a compelling solution that we have combined into a single HDK but our channel partners and distributors are excited to take to market. Completing the package is the cross leverage we are building between hardware, software and IP and the improving scalability of our processes. The term industry 4.0 dates back to 20 11.
There has clearly been progress towards realizing the industry 4.0 vision in cloud Computing, but not to the Azure endpoint. This is due telegically positioned to fulfill that pent up demand. Edge and Endpoint represents the vast market with hundreds, if not thousands of potential customers, that together represent very substantial aggregate volume potential. To address these widely varied markets and customers, We are integrating the reach of channel partners and distributors in North America, Japan, Taiwan, South Korea and select European countries. With this and our other initiatives, I believe we are well positioned to expand our market coverage and customer base beyond anything QuickLogic has realized in the past and deliver a balanced model of hardware embedded software that can be monetized, recurring SAS revenue streams, and IP that is able to deliver sustainable
Thank Our first question comes from Richard Shannon of Craig Hallum. Your line is open.
Brian, it's Steve. Thanks for taking my questions. A lot of great remarks, Brian, a lot creating a lot of questions. Congratulations on the on all your progress on QuickAI. Before I jump into that, Brian, just want to make sure that I heard you in your prepared remarks regarding seeing at least 50% total growth in sales in 2019.
Is that correct?
That's correct, Richard.
Okay. To dig into that a little bit further, you've talked about a number of initiatives, QuickAI stuff in hearables and wearables and consumer electronics devices and others. If you can give us a sense of where you expect relative growth from those categories to help us kind of think about how to model towards that 50% growth for the year?
Sure. So if we look at the 50% growth here, greater than 50% growth for the year, I would say probably like half of that is going to come from our EOS S3 in the applications that you were just referring to, which are that diversified set of in applications like smartphone hearable, consumer electronics wearables. If we double click on those categories, I would say that the hearables is probably the biggest singular area out of that, just because we've been working on that for so long and have quite a bit of the design wins in that area. The other ones are not far behind that though. I think the consumer electronics and we've already talked about as an established application adding voice, we see fairly substantial volume with that.
And then the fact that we're starting to ship products now for that 1st Japanese smartphone I'm hopeful that we'll be in probably 3 models this year of that. So that should start to kick in from the Q2 timeframe. So fairly balanced, but I think probably hearables would have the larger component of that for the EOS S3. If you look at the other new product categories that we've talked about, that would contributing to that greater than 50%. Obviously, the IP is 1.
QuickAI is 1 in SensiML would be another category. And I think from a modeling point of view, those would be fairly well evenly distributed at this point. I think as we get through the year more, We get a sense a little bit more about the SensiML business model and the customers that they have and see how the SaaS revenue is going to kick in. That may change. But I think for now, we're modeling all of those kind of split splitting the difference.
Did that answer your question?
Great. That helps a lot, Brian. It gives me a lot of context to think about some other pieces here. Quick, I obviously, as I recall, when you did that first introduction back I think in May last year you were thinking revenues towards the end latter part of this year and you've already seen some material revenues now. Wondering if you can I don't remember hearing anything about the scale of revenues reported in the fourth quarter?
If you can tell us that, but also just kind of a broader sense of the size of the pipeline, how fast people are moving through this? And maybe if you can get a sense of how big each engagement might be? Are we talking about units in the tens of 1000, 100 of 1000? Are you seeing any in the millions of units sort of thing? Any just broader or detailed view on that pipeline?
So a lot of questions in that one question, Richard. I'm going to try to answer them, but it's I don't, just remind me of the question I didn't answer. So, starting with Q4, it was definitely material, but not enough to break it out as a 10% item. It was not just one customer as I alluded to in the prepared remarks. I would say it's the first applications are more in the time series data.
Time series are things like motion sensors and microphones and applying AI to those for industrial type applications. The reason why I distinguished time series is because when we first launched in May of last year, we encompassed time series and vision. Some of these initial engagements we have that are moving faster is on the time series side. And I think that's probably A, because there's kind of demand and B, it's a less congested space. I mean, if you talk about AI, everybody that's doing AI generally is glomming onto the vision aspect of that.
And I think time series is sort of an reserved market at this point. Getting into the future, I think the distribution channel for us is really important. And that gets to your question about the types of customers. So the initial customers that we focused on for May were ones where we were largely doing the business development of that with partners because we didn't want to turn on our channel until we knew we had something that was going to be easy for the channel to sell. We had a sales conference worldwide 1 in Asia, specifically a few weeks ago where we turned on the channel to this And the benefit of using distribution channel is that they can now make a 10,000 unit a year customer a profitable customer for quick Whereas if we had to send our own folks on-site, obviously, revenue gets or ROI gets upside down pretty quick in that situation.
So with them wanting to take these 50 HDKs in Japan. And I'm sure there's going to be more as we roll these out with other distributors. We're going to start to see opportunities come in that are less than the typical $100,000 a year that we have the minimum cutoff for in our previous consumer market model. Means we're going to be able to serve a lot more customers. The ones that we've been directly calling on are still ones that are north of the $100,000 have engaged with some very large customers that are sort of the industry leaders in industry 4.0 and bringing predictive maintenance or condition monitoring into the industrial place.
There are big companies in there that have scale, not just in terms of how many customers they end up populating factories of but also they tied up to the cloud services. So, there's a lot of leverage there for us for the direct engagement. And then we'll use the distribution channels for the masses. By the way, that's one really key component around this HTK that I talked about in the prepared remarks. The starting price for that can be $500.
It's going to include a few months of the SaaS subscription for SensiML. That makes it so easy and frictionless for the channel to sell. This is really it's going to be a multiplicative effect on the channel coming opportunities. So it's actually pretty exciting.
No, I think you picked up most of them. So I will move on to different topics because that's some great detail there, Brian. Sune, maybe a couple of quick questions for you. Maybe on the guidance for the first quarter, your gross margin range a little bit higher than we've seen in the past or in the past quarter, although I understand there was an inventory reserve charge last quarter comparing those. But if you get a sense of directionally what's implied about product gross margins in there?
And also in licensing, will that be a flat or upper down sequentially in the first quarter?
Richard, thanks for the question. Yes. So, for Q1 guidance, we guided at the 55 percent plus or minus 3 percent for gross margin. This is based on a more mature product with a high growth margin. The other is with the QuickAI side, it carries much more, gross margin.
Now some portion of new product revenue also comes from SensiML SaaS revenue. So that helps with the gross margin. As we are moving as we're moving forward, more revenue coming from SaaS, from QuickAI and about the PGA, we just see our gross margins globally at that range 5%. Branch. So I hope I answered your question.
Yes, you did. One last quick one for you, Sue. I think you gave us a number of OpEx for this quarter. Is that a good range to think about going forward? Absent any big growth or is this kind of a baseline or is this not capturing the full run rate we should expect?
For the OpEx side. So because of the acquisition, so some more, we expect the R and D expense will be up, at the range that I guided $5,100,000 for Q1 2019, I think that's a good range to model it. I want to model it between, And about that range, 5,100,000 dollars, $5,100,000, $5,200,000, that range.
Okay, that's helpful. That's helpful.
I think I've asked enough questions. I will jump out of line. Maybe I'll jump back in, but thank you for all the detail. Appreciate it.
Sure. Thanks.
One moment please. Our next question comes from Suji Desilva of Roth Capital. Your line is open.
Hi, Brian. Hi, Sue. On the embedded FPGA opportunity, can you remind us when you go from an MTLA MTLA to a full license, what kind of bump up that is in a per customer license? And also, Brian, can you touch on, I thought it was per customer, but it seems like it's more per chip per project. So if you could clarify that, that'd be helpful.
Sure. So generally speaking, the MTLA, is just a support of maintenance agreement. So it's tens of 1000 of dollars. When it jumps up to an actual license where they're taking IP to embed into a chip, they're intending to take to production That's on the order of a few $100,000 or higher depending on the process node. And then to be clear to your second question, it's a per use license fee.
So every tape out that they do for a production chip would carry an individual license fee and then royalties thereafter one the actual device starts to ship in production.
Okay. That helps clarify that. And then on the quick AI customer who opted for the EOS instead in the design. Can you just walk us through the puts and takes of that decision and the dynamics there and why one customer go with the QuickAI module versus the EOS fallback. It's a lower ASP there, but I don't think that's the only element there.
So any color on the dynamics there would be helpful.
Yeah. So when we launched the QuickAI initiative in May, the vision partner that we launched with was nepesa with their NM500. And the specific engagement that we're talking about for Q4, was going down the path of using that nepes device which made our module with our device with an EPIS and EOS S3, a very compelling, easy to integrate solution. Then it have decided to change your mind and not going with the net percent of 500 instead of going with somebody else. And so we don't have any module at this point defined with that other company or other vision partners.
And so there was a pretty seamless transition for them to go back to EOS S3 ai as a device and then they would take on the responsibility of doing the hardware design with the other vision shipped by themselves. The other thing is that they are looking at using EOS S3AI for time series applications as well. And in that case, it would be on an entirely different ECB. And so there's no vision chip needed and therefore no module needed as well. So those are sort of the dynamics involved in that.
From a financial point of view, I mean, we talked about the module solution from us being mid double digits driving down to a single digit kind of number per you can see the impact that would have on overall revenue at the same kind of volume.
And just to be clear, a follow-up on that question, Is the module meant to be a production solution or is this transition from module to EOS S3 at high volume a natural, way that this would play out for a given customer?
It was originally intended as a demonstration vehicle in May. It became pretty clear that this one particular customer had value for just using that module as is. So we were actually productizing it to be a sellable production worthy module. And by the way, we're still looking at what modules would make sense for us to do. We have not abandoned that idea by any stretch of the imagination.
But we're going back and refining what that is so that we can actually make sure we define it correctly. And with the right ASPs. We're still targeting those much higher ASPs than the devices are.
Okay, great. And then the last question for me, and I'll pass it on the mature business, so you guided it higher in the first quarter, what's a more normalized run rate we should think about for the mature businesses in 2Q going forward. I'm presuming that 1Q is an elevated level or perhaps it is a new run rate. Any color there would be helpful.
Right. Hi, Suji. Thanks. So mature product, our model added about a 1,000,000 plus minus 10%, I would say, and that's probably at the high end. So $2,100,000 for Q1.
Our model at the 1 point Well, Justin, because this is a totally not under our control. It's like orders come in, we take it. There's no sales, marketing effort or any effort for the day. I'll be conservative. I'll model for the year at about one point $6,000,000 plus minus that level per quarter, per quarter, of
course. Yes. Okay. Great. Thanks guys.
Yep. Thanks.
Thank you. Our next question comes from Rick Meaton of Rivers Shore Investment. Your line is open.
Thanks Brian and Sue for taking my call. At CES, earlier this year. Did you have any other, customers other than the consumer electronics customer showing products with quick S3 in them?
Yes, absolutely. We did. We had a couple of hearable products. We had a new, how would I describe it? It's a head worn device, with our EOS S3 in it as well.
That was being shown in addition to the consumer electronics. Several. Those are the ones that come to
my mind. And these were being shown outside of your suite.
That's what I'm just getting at. They were being shot outside of our suite. Yes.
By the customer, okay. With the issues you had last year, involving close talk and all the other specifications on the move from push to talk to always on. Did some of the hearable products that you showed in January 28 team. Are you still working with those customers on redesigns for the new Amazon specifications and are most of these designs involving multiple microphones now?
Yeah. And I didn't get into that level on the prepared remarks, but it's worth bringing up here. So the designs that we had in the CES 2018 were all one microphone designs. And since then, we had announced a partnership with DSP Concepts, which has noise suppression and beamforming software multi microphone designs. And, it was good that we did that actually because the new specs from Amazon are really forcing everybody into a 2 microphone configuration.
This is nothing specific to QuickLogic. It's just the better performance you get with the 2 mic system. And so the majority of the customers that we were engaged with at that time in 2018 CS have either already changed or are already starting changes from one microphone to 2 microphone designs. So the ones that have changed, I think there's a couple now that have already done the change and they're going to be entering into the certification process. Others are having to go through that change now from a 1 mic to 2 mic configuration system.
It's not a big deal, but it still involves changing the hardware, adding the 2 mics and looking at from a system design point of view, how do you how do you put that pinhole for the microphone to get the sound from the industrial design?
Okay. So you have all of these particular design engagements that you continue to work on in 2019 that you didn't call out in detail in your prepared remarks. Is that right?
Yes, that's right. And by the way, the reason one of the primary reasons why we didn't call it as much detail in the prepared A, prepared remarks, obviously getting quite long for this call because it was a year end call. I want to go into more detail. The second is that there's been a few instances now where we try to be very forthcoming with information for investors, which I know is interesting and exciting. But it also leaves a lot of bread crumbs for our competitors to go chase.
And we've had a few instances now where that's caused some headaches for me. So We're trying to limit the names that we use on the call now, to ones where we don't view that to be a risk. And at the same time try to provide enough clarity for everybody on the call to understand the traction and be able to do the models that you guys need to do. But yeah, the short answer is there's other stuff that was at CS with other ODMs or IDHs that we've talked publicly about. That we decided not to include in the call specifically today.
Okay. So then Given that early given that in your prepared remarks, you mentioned that you had been working with JD dotcom for almost a year. And then finally announced the design win in the last few weeks. How many design wins are you working on right now? In, that involve EOS S3 and QuickAI.
And without getting into specifics on who they're about, how many things are in the funnel and can you provide any overall color about the nature or the average value of these wins or designs?
Yes, it's a good question. And I'll answer it. I actually have looked at the funnel. Obviously, periodically to understand where we are in progress what's so on. I would say that some of the macro trends, firstly, if we look at funnel from like 2017 and now funnel from 2018, the last 2 years separately, we're talking about design wins, definitely double digit design wins, like several double digit design wins.
And that spans primarily EOS S3 Obviously, some QuickAI and some EFPJ, but primarily EOS S3 because that's what we've been working on for the longest. The interesting note I would say is that in the last year, the hit rate has gone up like noticeably in terms of, I think, good targeting and having a good solid value proposition with the product that we have less losses as a percent of the total funnel, which is really good for efficiency for a smaller company. The second thing that I think is even more interesting is that I like to call this we're going big game hunting now in some sense for EOS S3. If you remember in the beginning, we were doing mainly business with Chinese IDHs and ODMs and that's sort of like bait fishing. The average size of the design is like $100 or less.
Now the average size of our design I just ran a report on actually is over $1,000,000 per design. So it's a pretty dramatic increase and I think that's a testament to the fact that we have a more complete solution. We have something actually very needed in the market now with voice for these bigger customers and we're getting much smarter about how we target and saying no to people that may want to use it that frankly we don't need their business. Does that answer your question on desirables?
Yes, that's very helpful. I appreciate the color on that. And thanks, Brian. That's all I have for now. Thank you.
Thanks, Rick.
Thank you. I'm showing no further questions this time. I'd like to turn the call back over to Brian Faith for closing remarks.
Thank you, Valerie. We'll be participating at the Embedded World 2019 Exhibition Conference in Nuremberg, Germany on February 26 through 28. And we look forward to potentially seeing some of you at this event. Our next conference call is scheduled for Wednesday, May 8th, at 2:30 pm Pacific Time. Thank you for your participation and continued support, and goodbye.
Thank you. Ladies and gentlemen, this does conclude today's conference. Thank you for your participation and have a wonderful day. You may all disconnect.