Hello, and welcome to the Everspin Technologies First Quarter Financial Results Conference Call. [Operator Instructions] Please note, this event is being recorded. I would now like to turn the conference over to Dave Allen. Mr. Allen, please go ahead.

Thank you, Keith, and thanks to all of you for joining Everspin's First Quarter 2017 Financial Results Conference Call. Before we begin the call, I want to remind you that this conference call contains forward-looking statements regarding future events, including but not limited to, our expectations for Everspin's future business, financial performance and goals, customer and industry adoption of MRAM technology in successfully bringing to market and manufacturing products in Everspin's design pipeline and the execution of its business plan and delivering stockholders an attractive opportunity with a long-term growth model and healthy profits. These forward-looking statements are based on estimates, judgments, current trends and market conditions that involve risks and uncertainties that may cause actual results to differ materially from those contained in the forward-looking statements. I would encourage you to review our SEC filings, including Everspin's financial results press release issued on May 12, 2017, Everspin's annual report on Form 10-K filed with the SEC on March 29, 2017, and other SEC filings made from time to time in which we may discuss risk factors associated with investing in Everspin. All forward-looking statements are made as of the date of this call and, except as required by law, we do not intend to update this information. In today's call, we will be referencing GAAP and adjusted EBITDA numbers. Adjusted EBITDA numbers are provided to enhance the investors' understanding of Everspin's operating performance as it primarily excludes certain noncash charges for depreciation and amortization, stock-based compensation expense and compensation expense related to divesting of common stock held by GLOBALFOUNDRIES resulting from our joint development agreement. The use of adjusted EBITDA is not meant to be a substitute for the results presented in accordance with GAAP but rather should be evaluated in conjunction with GAAP. The conference call will -- this conference call would be available for audio replay in the Investor Relations section of Everspin's website, www.everspin.com. Joining me today are Everspin's CEO, Phillip LoPresti; and CFO, Jeff Winzeler. And I'd like to talk -- turn the call over to Phill. Phill?

Good morning, everyone. For those of you who might not have seen our financial results press release, you can find the press release and an updated investor slide deck on our website. Before I ask Jeff to discuss our Q1 results in detail and provide our Q2 guidance, I would like to quickly make a few comments about our first quarter results as well as other recent activities crucial to our future business. First and foremost, I am encouraged by the product development execution by the Everspin team and the solid Q1 bookings for our Gen 1 Toggle products. Our design activity for both our Gen 1 Toggle and Gen 3 Spin Torque MRAM solutions continue to gain traction as we prepare for the production launch of our 256 megabits Spin Torque product. Revenue for Q1 was $7.9 million, ahead of our guidance of $7.3 million to $7.6 million, and up from $7.1 million in the prior quarter. Q1 revenue was driven by a mix of Gen 1 Toggle MRAM shipments, legacy products and nonrecurring engineering, or NRE, services. Our gross profit was $4.2 million resulting in a gross margin of 54%. Our EPS loss of $0.49 per share was in line with our guidance. Design activity for our Gen 1 Toggle MRAM products remained robust with the continued strength in higher density, higher ASP industrial applications. We also saw follow-on business with our enterprise storage and automotive customers for our Toggle solutions. The expected dollar value of our Q1 design wins was 34% higher than the prior quarter. Our next Gen 3 Spin Torque 256 megabit and 1 gigabit MRAM product development work continued to proceed on schedule. Last quarter, we reported 3 design-ins, including one with a major OEM for a flash array product. I am pleased to…

Thank you, Phill, and good morning, everyone. First, let's review our Q1 2017 income statement. Revenue in the first quarter was $7.9 million, with product sales representing 92% of total revenue or $7.2 million, while licensing, NRE and royalty contributed $660,000 in the quarter. Our Q1 2017 revenue at $7.9 million was up $816,000 or 12% higher than our revenue in Q4 2016. Looking at product sales alone, our revenue was $7.2 million compared to $6.9 million in product sales in Q4 of 2016. Gen 1 Toggle sales, which were down $174,000 or 3% from Q4 2016, were offset by our legacy products, which increased by $488,000 or 43% when compared to the prior quarter. Our gross profit for Q1 2017 was $4.2 million, an increase of $984,000 or 30% over Q4 2016. The resulting gross margin for Q1 2017 was 54% versus 46% in the prior quarter. Our gross margin expansion in Q1 2017 when compared to the prior quarter was a result of better yields and lower cost of our Gen 1 MRAM products, coupled with additional licensing, NRE and royalty revenues in Q1. Our Q1 2017 operating expenses were $10.1 million compared to $8.5 million in Q4 2016, an increase of $1.6 million. Research and development expenses in Q1 were $6.4 million versus $4.9 million in Q4 2016. The increase in Q1 2017 R&D expenses was a result of variable costs associated with our joint development efforts on Spin Torque process technology. SG&A spending for Q1 2017 was $3.7 million compared to $3.6 million in the previous quarter. The $100,000 increase was due to professional fees related to our year-end audit. Interest expenses for Q1 2017 were $230,000 compared to $365,000 in Q4 of 2016. Other income was $19,000 in Q1 2017 versus $346,000 in the previous…

Thank you, Jeff. Overall, I'm pleased with our business performance. Demand, as measured by our book-to-bill metric, continues to be strong, and execution of our business plan continues to remain on track. Key to our future success, our Gen 3 Spin Torque 256 megabit and 1-gigabit MRAM product development work continues to proceed and, we believe, can significantly expand our addressable markets starting in the second half of this year. Our design-win momentum continues in our targeted market areas. Our Gen 1 Toggle products, which target persistent SRAM applications with memory densities ranging from 128 kbits up to 16 megabits remains the bulk of our business. As I mentioned in my introductory remarks, we continue to see steady design-win activity in this area, including an increase of 16-megabit design wins in the industrial space that will gradually shift the unit mix to higher ASP products. In fact, nearly 3/4 of our Q1 design wins were industrial while enterprise and automotive accounted for the balance of Toggle design wins in the quarter. Let me share with you a recent industrial design-win example with a major factory automation supplier that serves a variety of end markets from aerospace to woodworking. Our Gen 1 Toggle MRAM is being used by this customer to store critical manufacturing and machine setup information in their computerized, numerical controller for high-precision machining of molds, tools and dyes. With our nonvolatile MRAM attributes, this customer now has a battery-free solution, which improves reliability, system uptime, and total cost of ownership. This design win will begin ramping later this year and we expect will remain in production for several years. Another important Toggle win in the first quarter was with a major auto manufacturer who will use our MRAM in a new emergency call system and high-end AVN, or…

[Operator Instructions] And today's first question comes from Kevin Cassidy with Stifel.

And also, congratulations on hiring Annie and putting on the 2 new Board of Directors.

Thank you.

Thank you, Kevin.

Yes, in your revenue guidance for the second quarter, can you give a little description of what the mix would be in products, legacy Gen 1 versus Gen 3?

Yes. Well, it's continued growth in our Gen 1 MRAM that's driving the majority of where we're going to see our product revenue grow. Legacy will continue to run at about the same levels it's been running over the last couple of quarters. We're expecting that we'll also have some additional revenue in the royalty and NRE lines like we did in this quarter. So in general, I think the growth comes primarily from our Gen 1 MRAM products and then, I think, our legacy and our NRE revenues will remain relatively flat.

Okay. And are you expecting some of these design wins would ramp in the production in the third or fourth quarter? You don't have to give exact numbers, but is that the trend we should be seeing?

Well, yes. Any design win that we get in Q1, we would expect to start ramping into production in Q2 and Q3, certainly.

Okay. And one other question. On the R&D spending being up, can you give a little more description of what that joint development cost were? And are they expected to extend into the third and fourth quarter?

Yes, the big variable spending in the joint development agreement, it's all consumable spending. So the major drivers in that are test wafers that we run for shaking out the process as well as additional mass sets whenever we do a rev of the product. In Q1, specifically, we had a couple of revs of mass sets, which are relatively expensive, and that's really what drove Q1 -- the Q1 number up from what we saw in Q4.

And Kevin, this is Phillip, if I may add as well. In the effort of expediting our 1-gigabit development as we have customers that are very excited to get that sample in hand, we are, obviously, doing more work in a compressed amount of time to get the part out quicker so that we can start that sampling process, and that dovetails into the answer that Jeff just gave you.

And the next question comes from Logan Bender with Canaccord Genuity.

Hey, guys. I don't know if you can hear me. This is Matt Ramsay from Canaccord. I was encouraged, Phill, to hear about another storage design win that you guys announced here on the call. And I know that we've been through some of these details in the past, but it would be, I think, really helpful for us to understand a little bit more the market opportunity regardless of exactly which quarter in the next few that some of these design wins ramp into revenue, and each of the larger customers will go at their own pace. But just to understand as these are fully ramped design wins in 3, 4 quarters from now what the potential is in terms of revenue for the storage design wins that we're talking about here whether it's number of units, price per MRAM cell, those kind of metrics just to help frame some of these opportunities. I know it's hard to do exactly but that, just to get an idea of magnitude would be really helpful for us.

Sure, Matt. So the 2 design wins that we've framed out with flash array storage providers are actually really pivotal from our perspective. These are use cases that we specifically optimized our product to address, and obviously we've been working alongside with many of these storage companies over the course of the past year or so. Some of them sampling our 256 in the very early days, in August of 2016. What I can say is the typical architecture for these flash arrays are that they would use a, let's say, most likely, a U.2 form factor drive. And these drives are typically put into a chassis. Many of them can go up to 24 of these drives in the chassis. The use of our Spin Torque MRAM would be as a buffer or cache for each of the U.2 drives, and in most cases, we're expecting to have at least 5 of our MRAMs per U.2 drive. In some cases, depending on the customer's configuration, it could have as many as 9 of our MRAMs in that U.2 form factor. So to kind of give perspective, if they populate the chassis with 24 of the U.2 drives, you're -- you would multiply 5 or 9 times the 24, and that would give you a perspective of how many MRAM units can go out with a particular chassis from one of these suppliers. Obviously, the unit volume of the chassis that they ship are going to vary with the performance rating and the density of the specific flash array products. But you can kind of frame up those types of volumes by looking at the leaders that are out there today and what they typically ship per quarter or per year, and, obviously, being designed into these 2 with pretty significant players with advantages and now having MRAM as another advantage, we're hoping and expecting that they'll get very good traction for these products. Again, the predictability of the launch dates is always really up to the customer. But the fact that we quickly went into a design-in with a major OEM and that very quickly has moved to design win, they're placing orders now for initial builds and the value -- and testing, is very good progress that these -- that this company wants to move fairly quickly. And, of course, getting a follow-on is also significant for us because, again, it proves the value proposition of our solution.

Phill, that's really helpful. And just a couple more questions from me. I think one is just how should we think about the revenue inflection of the nvNITRO card? And you talked about some traction with some of the high-frequency trading applications, and I can imagine some other applications that might have similar application dynamics. I guess that my understanding is that we should see some revenue here relatively quickly, and I'm just trying to get a gauge of how big of an opportunity that could be and the timing of that relative to some of the longer-term design wins you just discussed.

Sure. So nvNITRO offers an opportunity to accelerate the MRAM product into the market because our focus is to have a prequalified board or module. In this case, it's a PCIe card. In the future, we're looking at M.2 and U.2 options, which we discussed at the last call. The process you have to go through, there's 2 particular channels that we're focusing on. One is with the providers to these high-frequency traders that integrate the systems together and that are commonly supporting them. So the cards that we're having evaluated are going into some of the channel that supports that effort, and they're obviously doing performance metrics, they're comparing workloads, different workload situations to identify the used cases that are specifically get an advantage by using our nvNITRO. And all the data so far has been very encouraging. There's several interest points, and we expect that that's going to expand over time. It's hard to quantify that ramp and the particular number because we really don't have a clear definition of where it's going to go yet, although the interest level is extremely high. The second path would be with a major server OEM, for example. Many of them provide an option card program so that typically if you go to one of these big server or storage providers, you can build a system to your needs, and Everspin's in the process now of working with a couple of these big server storage providers to enter and get our nvNITRO cards qualified into that option program, which would then enable -- people that go to that OEM's website to configure a system and select a qualified nvNITRO card into that server. So again, that is -- the first step, you have to get into the program, get on the website. But once we're in that, then obviously, it opens it up, and it's hard to, again, define what that actual number can be. The first step is enabling in both of these cases, and that's what we're doing through the sampling right now.

Got it, Phill. That's really helpful. And then, just lastly, if you'll indulge me, the question that I get most often from investors is this is, obviously, a breakthrough in new memory technology, and you guys are on your commercializing your third generation now. But there are some competitors in the landscape we continue to hear a little bit about, MRAM work that's going on at some of the other foundries in addition to your partner, GLOBALFOUNDRIES. Maybe you could just walk us through briefly, if you could, the -- how far you guys are -- feel like you are ahead of competitors in productizing and manufacturing MRAM at scale? And how we should think about the competitive landscape over the next number of years?

Sure. Thank you, Matt. So the -- obviously, our competitors don't communicate their road map and their schedules and their plans in any length of detail, so I really can't comment specifically on where they're positioned. But I can say from our perspective, having our partner, GLOBALFOUNDRIES, up, we've been working with them since late 2014, we transferred a already working process to their line, upgraded to 300-millimeter tools, we then are -- optimized that process and were able to implement our process on our 40-nanometer, 256-megabit chip and actually bring it into production ahead of the schedule that we were projecting by qualifying it in early March. So from that perspective, I think that gives our foundry partner, GLOBALFOUNDRIES, an advantage because they're already qualifying a production process on their line, and we're clearly going to be starting material to address not only sampling requirements for 256 and potential production requirements, but we're also now moving our 28-nanometer 1 gig through that process and getting that qualified as well. The announcements that we've -- that we understand to date that are out there are from foundries that compete with GLOBALFOUNDRIES, and their target, at least what we can get from the press releases that have been published, is that they're planning to use the MRAM technology they're developing in-house for embedded memory solutions that they provide to their foundry customers. So this is, what we believe, completely validates what we have been targeting from the very beginning and why GLOBALFOUNDRIES is working with us is that there is a big advantage to having MRAM as an embedded solution in these small geometry processes where it's very difficult to scale flash technology. And so as you see, process nodes of 28 and below, you don't commonly see embedded flash available. And so we believe that gives an advantage and thus, it's only logical that GLOBALFOUNDRIES' competitors are going to try to move to compete with them since they have that advantage from us, being the only company that has brought any MRAM to production and the fact that we're already going into production on their line. So I think there's an advantage there. It's hard to quantify it because we don't know where the others are, but it doesn't come as any surprise to us that the global's -- GLOBALFOUNDRIES' competitors are moving to try to compete in that space.

And the next question comes from Rajvindra Gill with Needham & Company.

Just a point of clarification, just so I have it. So coming into the quarter, you had 3 design-ins and one of which is now transitioning to a design win going into production. And then you added another design-in. So in total, you have 4 design-ins, one of which is going into a design-win phase. And I think you mentioned about the other 2 design-ins. Can you just clarify that as well for me?

Yes. So from what I can get from what you said, I think you got it correct. In total, there have been 4 design-ins. Two are with the flash array architectures that we've talked about. One was with an M.2. We talked about that last quarter. And the other is with a U.2 SSD. One of them has moved from design-in to design win, and what that means for us is to register it as a win, it means they're buying or they placed orders now for product quantities to allow them to start doing test evaluation and finalize their product. So that's what one of these -- and then in this case, it's the flash array solution. Also, out of the previous 3 that we announced, one of them has actually moved into the production phase, which means they've already put parts in, they've qualified a product and we're now waiting for them to place their initial production builds. And the other one that we announced last quarter is now preparing to do the same thing, place orders to start their initial NPI builds, for example.

And can you give us a rough time line from design-in to production phase to design win? I know it's going to vary by customer...

Yes, Raji, it varies quite significantly. And you could see in this case that in March, when we had our earnings call, we were able to register an in, and in a very short amount of time, that customer has already moved to the point where it's a win. And that happened in, let's say, less than 2 months. What we typically see is that this could be a 3- or 4-month cycle for these conversions, but we believe, this is something that we've communicated during our roadshow, because we've been seeding the market with samples all the way back into August of 2016, a lot of the customers we're engaged with have already been preparing, testing, evaluating. Some cases, they had their production board already laid out, and they were just waiting for us to provide the green light that the part was going to go into production. And so some of those are actually in a position to move really quickly. Whether they choose to do that or not is not necessarily in our control, but I'm hoping and we feel that some of these other companies we're engaged with will move as quickly as the 2 or 3 that we've talked about, and hopefully, we can move them into production status in the next quarter.

And you gave a very excellent example of trying to frame the opportunity with respect to this U.2 form factor, whether it could be 24 U.2 drives using MRAM, either 5 MRAMs up to 9 MRAMs per U.2 device -- U.2 drive, sorry. So just to quantify it a little bit further, if we're talking about 5 to 9 MRAM per U.2 drive, I believe the ASPs you had mentioned were in the range of $5 to $10. They might be a bit higher for the SD MRAM. But if I just assume $5 to $10, we're talking about $600 to $2,200 per system. Is that kind of the -- I'm just doing the basic math, I'm looking at nearly $600 to $2,200 per system of MRAM content.

Yes. So I believe your $5 to $10 range is off. It's going to be higher than that for the initial launch of our product. And so I would say that, that's definitely lower than what we're going to see based on our pricing today. So yes, I think that model of $5 to $10 is off. It's going to be at least 2x or somewhere in that range higher than that.

Okay. So we're looking at then something closer to $1,200 to $1,800 or even up to $2,000 per device?

Per chassis, if the chassis is populated with 5 or 9 and they have...

24 drives.

Right. 24 drives, right.

So actually, it could be $1,800 to $3,200 per chassis?

That's correct.

Okay. Got it. That's helpful to frame that. And the next question I had was on the ADAS, the Advanced Driver Assistance Systems, representing a key growth area for the 256-megabit product. I know you had mentioned that you were expanding your scope with one of your large automotive OEMs with plans to offer additional Gen 1 but also Gen 2 MRAM products. Can you talk a little bit about your view of the Gen 3 product? How much of it -- how fast is it gaining traction to support ADAS, Advanced Driver Assistance Systems? What are some of the applications that are conducive for an MRAM memory architecture?

Sure. So today, we're selling into the automotive space our Toggle Gen 1 product. Its primary function is a data logger. It's collecting data real-time from various sensors in the car system or accumulating data for decisions being made by the microcontrollers. There's upwards of 60 microcontrollers in vehicles these days, and that's growing. So we continue to see in some of these use cases where the density requirement really doesn't exceed our Toggle densities, which go up to 16 megabit. And so in some of these applications, they can continue to use the Gen 1. However, some of the new features that are being discussed and obviously, they're in test stages and expected to grow, maybe launch in vehicles in 2019 or 2020 and beyond, these vehicles are now, in addition to increasing processors, increasing the sensors that they're -- the number of sensors that they're using, they're also increasing the number of cameras that are either external or internal to the vehicle for tracking various types of information. Now depending on who you talk to or the way they want to craft the product, they may want to record. Some manufacturers are thinking of recording 30 seconds to 1 minute of video and refreshing that real-time so that you're rewriting into a memory that often while the vehicle is in use. And so collecting that, they obviously need a high-bandwidth interface, perhaps a DDR-type of performance. It could be a high-speed interface like a quad spy or octal spy. Those interfaces, again, are available for our product. And there's other applications where I'm reading that they might even want to record and store that up to an hour's worth of information. So the requirements vary quite a bit. But the advantage that we offer them is this high-bandwidth,…

And just last question. I mean, in terms of how long you're going to record the data -- the video data, is that decision going to be made by the insurance companies? By the OEMs? I mean, what's been your finding so far?

Yes. There is no standard right now, so it's somewhat of a free for all. But I have been reading that, obviously, insurance companies or various agencies that are going to mandate these let's call them for now like black boxes, they're going to have to set some kind of standards as to how much data and how long it has to retain it. But as I said, right now, it's kind of across the map today. We even have some applications that they're taking snapshots like literally photos and storing them as supposed to real-time video. Even though it's a camera, they just take photos and snapshots, and they can do that in like, let's say, a 16-megabit Gen 1 part. And others are looking at really high-definition, high-quality video for some of these ADAS systems where that increases the memory requirements significantly. And then, obviously, they have trade-offs to do that. If they want a higher definition video, they have to put more memory in there and perhaps they may only store shorter amount of video time as opposed to taking lower-quality video and taking a longer snapshot, minutes or even longer. So again, no standard, everyone's got different opinions and different approaches, and we're trying to work with all of them to see how we can address the need.

And the next question comes from Richard Shannon with Craig-Hallum Capital Group.

Listen, maybe I'll ask a question on the guidance, revenue guidance, for the second quarter. I just want to verify here. It sounds like most of the growth you're seeing sequentially is coming from Gen 1. Is there any material amount of Gen 3 revenues in there, even from a sampling or preproduction phase there?

Yes. So I think we've been pretty consistent that from a ramp perspective, we just introduced our Gen 3 product for production at the end of Q1. That's going to take some time for people to design it into their systems and start selling it. And then, obviously, on nvNITRO, which would be the other Gen 3 product that we would see revenue on, that product, again, is not going to be productized until the end of this quarter. So the opportunity to see significant Gen 3 MRAM revenues in Q2 is pretty slight.

Okay, that's what I thought.

And Richard, just to add to this. The designs that we have quantified today and in previous calls are just going to production status or are just going to design win status, so you can imagine that, that doesn't really define any significant number. We're pretty consistent that we believe these will start moving in the second half, Q3 and Q4. And so at least these 4 are progressing at the rate that we feel that, that's still obtainable.

Okay, great. That's what I thought. I just want to make sure. So the growth we're seeing from the Gen 1 products, I wonder if you can help us understand the drivers here. You talked about a few design wins. It sounded like they'll probably start hitting after this quarter. So I'm wondering by end market or breadth of the increase you're seeing in the second quarter, it'd be great to hear as well, please.

Yes. So the Gen 1 product, we've highlighted and characterized the automotive opportunities where our existing customer in that space has adopted our part in more vehicle types. These platforms that they have could have various different models and different levels of their audiovisual navigation systems. And so what we're seeing is them converting from an existing solution, more and more of those platforms and specific models to an MRAM-based solution, which is great for us. So we're seeing that increase in the automotive side. As Jeff highlighted, the mix impact, based on projections that we've had from this particular manufacturer, they're running ahead of the numbers we initially expected because of this adoption into more platforms. So I would say the automotive is driving some of the strength in the units. But also, we've had design wins in the industrial space that they take a little time to gestate in the industrial space, but they last quite a long time. And even when you move Gen 1 industrial design win into production, it could take them 3, 6, even 9 months before they get to any appreciable volume. So some of the design wins and ramp to production that started last year, we're now seeing those higher densities, 16 meg, ramping harder and into higher volumes, and that's driving some of the growth as well.

Okay, great. I appreciate that characterization. Great to see. So let me ask you about Gen 3 engagements with the storage in flash array market you talked about, a few design ins and design wins. We're curious if you can kind of give us a broader view on that environment here. There's obviously a number of OEMs, both large ones as well as some start-ups out there that are targeting the flash array market. I wonder if you can characterize the breadth of your engagements to the extent of what you're seeing across the market there. And why wouldn't someone want to use your device as opposed to more mainstream, historical usage of backups that they've used in the past?

Sure. So we're obviously heavily and actively engaged with virtually all of them, whether they're established players or some of the new start-ups that continue to pop up with various different advantages. But one of the clear trends, and this isn't new to the electronics industry, is that in form factors are getting smaller or definitely not getting larger, but at the same time, they're trying to pack more density into these form factors. And what happens in a lot of these cases, let's say, in a U.2 form factor that goes into a flash array chassis, just think about that U.2 form factor. It's an enclosed form factor. The size is fixed. It's got to plug into the chassis, so you can't expand it. And you're now packing more and more NAND in there because some of these U.2 drives are going to 10 terabytes, and some manufacturers are talking even 2x or 3x that density in a U.2 drive. Obviously, along with the higher density, they have to deliver better performance. So when you put more into that form factor and you run your controller at a higher performance, some of them are using FPGAs and some of them are dedicated ASICs, the heat that's generated in the power that's generated to service that inside that form factor is rapidly increasing. And with that, if you choose to use supercaps or capacitors, you actually have a kind of a double whammy of a problem. One, the capacitors take up space in that small form factor, which negates your ability to add more NAND and increase density. But even worse, if you do have the caps in there, you have a thermal problem that you'd have to contend with, because these caps and supercaps, as we all know, their…

Okay. Great. That's great characterization, Phill. My last question, I know you mentioned the time frame for sampling. I missed that. And maybe, Phill, if you can expand on that topic, on the 1-gig development. I wonder if you can repeat what are the time frame you expect to first sample the device. And then, what's the earliest time frame you'd expect the customers to start putting in a design -- in getting into that design-in process?

So we're aggressively working right now. As we characterized, some of the Q1 expenses were higher because we're doing more work in parallel to try to get the cycles of learnings and get our 1 gigabit in a position where we could start sampling it. There's various stages of sampling. We -- one is an engineering sample level, another one is a commercial sample level, which means it's getting ready to transition to production. So at this point, we're trying to at least get some of our customers' parts late -- before the end of this first half, let's say, perhaps early in the second half of this year so that they can start planning out their systems, laying out their boards and evaluating the 1 gigabit. So this is fairly consistent with what we were expecting to do. Obviously, getting it into production is going to take a little bit more time, so we're expecting that to happen later in the year. But we're making good progress in-house. We've obviously brought in various different material grades, and the testing in-house is going well, and we hope to get those samples into some of these customers maybe late this half, early next half.

And that's all the time we've allotted for questions this morning. So at this time, I would like to return the call to Dave Allen for any closing comments.

Thank you, Keith. Everspin will be participating in 2 investment conferences this quarter: The Stifel 2017 Technology, Internet and Media Conference on Monday, June 5 in San Francisco; and Needham's Inaugural Automotive Tech Day on Tuesday, June 6 in New York City. We look forward to seeing many of you at these events. Details for these conferences will be available on the Investor Relations section of everspin.com. Thank you for your interest in Everspin, and goodbye for now.

Thank you. The conference has now concluded. Thank you for attending today's presentation. You may now disconnect.