Cirrus Logic Inc (CRUS) 2017 Q4 法說會逐字稿

Ladies and gentlemen, thank you for standing by. Welcome to the Cirrus Logic Fourth Quarter and Full Fiscal Year 2017 Financial Results Q&A session. (Operator Instructions) As a reminder, this conference call is being recorded for replay purposes. I would now like to turn the conference call over to Mr. Thurman Case, Chief Financial Officer. Mr. Case, you may begin.

Thank you, and good afternoon. Joining me on today's call is Jason Rhode, Cirrus Logic's President and Chief Executive Officer; and Chelsea Heffernan, our Director of Investor Relations.

Today, we announced our financial results for the fourth quarter and full year 2017 at approximately 4:00 p.m. Eastern. The Shareholder Letter discussing our financial results, the earnings press release, including a reconciliation of non-GAAP financial information to the most directly comparable GAAP information, along with the webcast of this Q&A session, are all available at the company's Investor Relations website at investor.cirrus.com. This call will feature questions from analysts covering our company as well as questions submitted to us via e-mail at investor.relations@cirrus.com.

Please note that during this session, we may make projections and other forward-looking statements that are subject to risks and uncertainties that may cause actual results to differ materially from projections. By providing this information, the company undertakes no obligation to update or revise any projections or forward-looking statements, whether as a result of new developments or otherwise. Please refer to the press release issued today, which is available on the Cirrus Logic website, and the latest Form 10-K and 10-Q as well as other corporate filings made with the Security and Exchange Commission for additional discussion of risk factors that could cause actual results to differ materially from current expectations.

Now I would like to turn the call over to Jason.

Thank you, Thurman. Before we begin taking questions, I would like to make a few comments. For a detailed account of our financial results, please read the Shareholder Letter posted on our Investor Relations website. FY '17 was a phenomenal year for Cirrus Logic. We delivered 32% year-over-year revenue growth and expanded our operating profit and earnings per share. We significantly broadened our product portfolio to target many of the emerging markets that we anticipate will spur opportunities going forward. As demand for differentiated audio and voice experience continues to proliferate across a range of products and tiers, our ability to provide best-in-class components to span the complete audio signal chain has placed the company at a competitive advantage. We remain focused on growing our share with new and existing customers, including increasing our penetration of flagship devices, while pushing certain features and functionality in the mid-tier phones. With the portfolio of components and an extensive product road map, we expect to gain momentum selling a range of smartphone and digital headset smart codecs, boosted amplifiers and hi-fi decks into a variety of Android OEMs as we move into FY '18 and beyond. Longer term, we continue to invest in MEMS microphones and are excited about potential opportunities in adjacent markets, including smart home and new technologies, such as voice biometrics, as we believe they will further expand our served available market.

With interest in high-performance ultra low-power audio and voice products accelerating across a wide range of device categories, Cirrus Logic is well positioned to capitalize on these opportunities. The company has a comprehensive product portfolio that has been tailored to address the various price and performance requirements of these markets and outstanding, engineering and technical marketing and supply chain teams that are focused on executing our innovative road map, which we believe is essential for future growth.

Before we begin the Q&A, I would also like to note that while we understand there is intense interest related to our largest customer in accordance with our policy, we do not discuss specifics about our business relationship.

Operator, we are now ready to take questions.

(Operator Instructions) Your first question comes from the line of Matt Ramsay with Canaccord Genuity.

Jason, there is no, I guess, indirect way to ask the question and I appreciate the sensitivity, but the question I think many of us are getting the most often and regardless of customer is, many smartphone OEMs you've seen over time try to do more integrated functionality themselves or design certain products for vertical integration or do themselves, that includes many of the OEMs not just your largest customer, so -- and like I said that's the question we'd be getting most often, lots of things going on in the market unrelated to your business. But I wonder if maybe you could talk a bit about how important it is, the functionality that you provide? And those -- and that functionality be provided in discreet components to your customer base, regardless of the size of the customer?

Sure, yes, it's, to some degree, part of the name of the game in being the mix signal peripheral supplier. And there are some years, in some cases, where we stay ahead of that curve and continue to add digital functionality. In other years, and we have seen a case of that this last year with our -- just recently, with our second largest customer, where some of what we had been able to provide previously was integrated. The best defense against that, and I think it's a real good one, is real close coupling with -- between the analog and the digital of our devices. So if you take, for example, noise canceling of the sort that we provide in the handsets, the analog and the digital in that particular case are really closely coupled. It's a very high bandwidth, very low latency, very low-power connection. And to fracture that across a multichip boundary would be a really clumsy system partition. It's not to say it couldn't happen, but it would almost certainly consume more power. It would probably be more expensive. And so when you've already got something like that working, customers tend not to mess with it too much, and we expect that to be the case for the foreseeable future. On other cases, some of the Android components, part of our strategy is providing programmable DSP capability where features that are still in the process of being matured can be implemented on our device until they have been matured, and then sometimes they can get integrated into the AP or somewhere else that proves to be more cost effective or, in some cases, is perceived to be more cost effective, but in fact, isn't. So generally the defense mechanism of differentiating around the analog and mixed signal capabilities, certainly, things like the amplifiers where much higher voltages 8, 9, 10 volts and growing, you're not going to see that integrated into an SoC for the foreseeable future, I mean -- and you would just make it more expensive and riskier in the process. So we see that from time to time. Overall, we see good opportunity to continue to grow the content we provide into handsets -- life is not a straight line up and to the right, and this year and any another year is no different, but we see really good opportunity over time to continue to grow the business. And I don't see anything unusual that's happening in -- along the lines of the question you have asked. It feels like just -- that's life in our business.

I appreciate it. I'm relaying a question there, so don't shoot the messenger, but...

No problem at all. It's a good question.

I think as a follow-up question, one of the things that was maybe a little bit disappointing entering this fiscal year was the speed at which digital headsets were able to be brought to market and adopted outside of your largest customer and the third parties to support the new lightning ported headsets at Apple. Maybe you could talk about the work that you have done with other OEMs and with some of the third-party headset providers. And how that's progressed, sort of a status update of that part of the business would be helpful.

Sure. I think we're in a really great position to continue to capitalize on a very large market converting over to digital connectivity. I think it's a little bit the nature of the beast that's outside of our largest customer, that the market is pretty fragmented, and as such, it generally moves a little slower than things would in a well-ordered world. But nonetheless, we see it happening, we see more movement towards USB-C connected headsets, every day, we see a lot of interest in that. We have got some exciting products that we're working on that we think will help accelerate that over time, and so more on that as things unfold. But we remain very excited about that market, and I think we're better positioned than anybody to capitalize on it.

Your next question comes from Tore Svanberg with Stifel.

A few questions, I'll start with the sort of long-term markets, first. You mentioned in your Shareholder Letter that you're collaborating with a market-leading customer sort of as a reference design for smart homes. So just wondering if you could elaborate a little bit more on that. And should we think of this as may be starting to generate revenues sooner rather than later?

Well, I mean, we don't like to preannounce any of our customer's products or things that are available, but in due time, that's something that I expect we'll do some press around, so we'll take the cloak off of that. But the smart home and taking the technologies, a lot of them that we've already developed and really gotten pretty close to perfection for the handset space are real natural fits in the smart home that's obviously a product category that's getting a ton of interest. I don't know that there is a real staggering number of units shipped today, but it's definitely a market that we're excited about and think is squarely in our wheelhouse. So we're working with the number of folks, including as we said, we view as the current market leader there and more on that later. But these are things like just making the voice interface work better and noise making -- things that are -- a lot of the products that are out there today in a connected home environment are wall-powered, which makes a lot of the signal processing significantly easier if you're able to throw a bunch of microphones and a lot more signal processing capability. But that same functionality is real desirable in a battery-enabled form factor or a variety of different battery-powered form factors. And so that's -- where hard signal processing meets ultra-low power and battery performance that's kind of where Cirrus has made our name and that's the kinds of things that we're focused on. So not a whole lot of more color to provide on that. I don't see that as a huge revenue opportunity for us, say, this fiscal year. But over the coming years, I think it's something that we'll continue to invest in, and ultimately, could be something pretty meaningful for the company.

Very good. And moving on to the portable audio segment. There is a lot of reference again in the Shareholder Letter to tiers of handsets. So it just seems like there is a real diversification effort going on. As we think about the growth for this year, is it fair to say that, that growth is probably going to be as diversified as it has been in a while?

Yes, I think that's only fair to say. We've -- it takes longer than people imagine to broaden the product line out to have devices that are applicable across the bulk of the Android customer base as well as applicable up and down the product range with any one of those customers might have. We made good progress starting to penetrate the mid-tier last year. I would say, we are making great progress now in penetrating both additional flagships in the China market, with things like our very high-performance hi-fi D/A Converter that we launched recently as well as making progress with really, really sticky products like the smart codec family and then rolling in the amplifiers as well. And those are with existing customers in the mid-tier as well as that's starting to add some new names to the portfolio. So I think we are plugged in with all of the right folks. We don't generally get into a whole lot of details about who exactly, but hopefully, the kind folks with their teardowns and whatnot will do some of that for us over the coming year. But I'm very excited about this: the position we are in, the progress we are making and the results the team has been delivering.

Your next question comes from Chris Rolland with Susquehanna.

You guys highlighted in the letter around the increase in OpEx in Q1. You said it was from employee and product development expenses and tapeouts. So I guess, my first question is, are those incremental development costs, are they for your largest customer, primarily? Or are they associated with the Android market, primarily? And then secondly, on the tapeout side, how much extra do you think was actually in there for extra tapeouts?

Yes, we really don't break any of that out. I just -- but it is -- I would say that the March quarter was notable in that it just -- that it was a relatively low quarter for the expensive tapeouts and that the June quarter is relatively high. That in a well-ordered world, if you were doing however many tapeouts over the course of the year that would be nice and evenly spread out as it happens, we've got a number of Rev-As of pretty significant products that are in. These more expensive nodes, 55-nanometer, that are out, that are kind of stacked up. Some of them could have gone in March, but ended up going out in April and so forth. So -- but it is indicative that we've built a pretty powerful R&D team. They are taping out new products that we are real excited about. And then just employment expense, there is a myriad increases in there, there is variety of other factors, but anyway, it all kind of added up to the guide we gave. And you can expect that, as we go forward, we will continue to add to R&D in a way that is sensible, relative to the top line growth of the company. We're trying to keep a real tight lid on the SG&A side of things, so that we can get the benefit of the leverage there. But we're not short of opportunities to invest in, and we're trying to make sure we do that prudently.

Okay. Great. And maybe kind of, on that note, perhaps you could remind us on the timing on some of the product transitions to new nodes that you are doing there. And also, on the gross margin side, you talked about supply chain efficiencies driving perhaps some upside there, maybe you could comment on that too.

Sure. I mean, I would expect 55-nanometer to work, be the work horse node for our smart codec product line for a number of years. It's a really, really good mix signal node, and additionally, it just happens to fit the tradeoff of analog and digital area very nicely. We are pushing them to 28-nanometer with the voice biometrics device that we are working on, which is reflective of the fact that, that's a lot of digital MIPS and memory that go into making that product work. And I would anticipate that, as we follow on that device with things where we have integrated that into a mix-signal codec that, that will drag the mix-signal stuff down into 28-nanometer probably sooner than in the process life cycle than we got to 55. But as it stands, if you took one of our existing smart codec's from 55-nanometer, migrated it to 28-nanometer, you might save a little bit of power, you would save some area, but it would definitely be more expensive. So it's not -- it's very much not the case that it's similar to the memory or the completely digital device domain where the way for pricing is such that if you migrate down, you almost always save cost. In our case, it almost invariably adds cost. So it needs to be enabling something that makes that worth it. Probably the biggest product transit -- node transition that's undergoing -- that we're undergoing right now is the transition of the amplifier product lines from 180-nanometer to now we've got a device out that we are out sampling and working on, that is in the 55-nanometer process. So that's coupled with a BCD kind of a back-end, because, again, these devices need 8, 9, 10 volts. So 55-nanometer, coupled with that kind of power and voltage by process, is a pretty unique proposition. We think that's going to give us a tremendous advantage versus the competitors out there. I think by the time bulk of the road map in the amplifier space is migrated over the 55-nanometer, it's going to make us very difficult to compete with. And that's a really good example of what I was just talking about. If we took the existing amplifiers that we've been doing and move them from 180 to 55, you are making some pretty extensive power transistors at that point. So it needs to be that we're bringing enough digital capability with it to make that value add be worth the additional cost. And I think we are doing that well, the reception of the products that we're outpitching and the road map that we're looking at going forward, I think, is definitely validating that approach.

Your next question comes from Blayne Curtis of Barclays.

This is Chris Hemmelgarn on for Blayne. I guess, I too will start with kind of a longer-term question. Yes, you've guided fiscal '18. As we look beyond that, though, I mean, you got a lot of irons in the fire, in terms of voice biometrics, active noise canceling. Historically, you've kind of had a nice little content uptick every couple of years with your codec. I guess, just as we look beyond this year, what do you see is the best opportunity to really drive growth? And what are you most excited about, as you look at everything you got in the portfolio?

Well, I mean, the premise of your question hit the nail on the head. We've absolutely got more irons in the fire than we've ever had. That remains a -- the focus of the company is making sure that we're keeping really tight reins on that, and we are not investing in too many things too quickly. But I feel very good about the balance of that across the board. If you look at -- we are day by day making great progress on the microphones. It continues to be a very difficult challenge, but the opportunity there to ship potentially in the hundreds of millions, if not billions of units, over time is there. And it’s something that we need to just keep -- the engineering team needs to just keep turning the crank and bringing out the supply chain side of things. And we are aligned with the right customers. I think, in general, they are pulling for us pretty hard, and that's a huge opportunity relative to the size of the company today. Voice biometrics, I don't see any reason that the opportunity is in at least as large as fingerprint, which is, again, large relative to the size of the company. Again, that's something that's got capital, our research left in it between now and the finish line and a lot of heavy lifting is still to be done. But it's -- the results are very promising, we are still on track to tapeout our first device this quarter. And a lot of customers have expressed a great deal of interest. Moving a lot closer in on the time horizon, though, as I said, we're making good progress, broadening out in additional flagship devices, we are making progress with both smart codecs and amplifiers, in some cases, more than 1 amplifier in mid-tier devices as well. There is an awful lot of units shipped there, though it's not going to be the same -- our expectation, on average, is it's not going to be the same sort of content that we have in a flagship. So -- but a smart codec for mid-tier device being on the order of $1 and $1.50, and amplifier $0.50 potentially 2 of those, I mean, and then you start looking at the headset attach as well, and it adds up to a pretty compelling opportunity as well. And then, of course, finally, the headset opportunity is -- as Android continues to make progress and switching over to various Android customers, either switching over to shipping the USB headset or a USB-C to headphone adapter, we see a lot of opportunities there. And kind of, as I mentioned early on, not a lot of details yet. We are working on some things that I think will make that transition a little easier for our customers and little more compelling user experience as well. So I can't think of anytime in the company's history, at least when I have been involved, that we've had as many things like that, that are large relative to the size of the company today. And in any portfolio, you expect you have some number of home runs and a few things that don't pan up like the way you expect, but with the scope and scale of things we are investing in today, I'm very comfortable with the markets we are serving and the opportunities we are going after that still leaves us meaningful growth opportunities in the coming years.

That's very helpful. I guess a quick follow-up. Perhaps more for Thurman than Jason. But I noticed you have been paying down your debt at a pretty healthy pace and it looks like, I mean, at the pace you've been on, you should be done with that within the next couple of quarters probably. How do you think about -- I mean, you've got pretty sizable cash flow, how do you think about using that after you finish paying off the debt? What are the main objectives. If you could talk about that, that would be helpful.

The main focus is -- I mean, we did a small acquisition this last quarter, and we'll still look at those types of opportunities that will push us forward from a technology standpoint. Larger ones that are actionable and are interesting to us are much harder to find. So -- but we will continue to look for those and try to take care of it that way. After that, we will be looking -- we continue to look at share repurchases and other investments along the R&D side. So on a opportunistic basis, we will look at the best way to utilize the cash going forward. And as we have noted, we have cash building up overseas, and we have to look at that challenge as a worldwide management of the cash. So...

Your next question comes from Charlie Anderson with Dougherty & Company.

I wanted to start, Thurman mentioned the acquisition and I saw some references to it in the Shareholder Letter, dealing with machine learning and neural network. So I wondered, Jason, maybe if you can just expand on that a little bit, sort of where that crosses over in terms of your competencies, and how that acquisition is going to help you? And then I got a follow-up.

Sure. Yes, I appreciate you highlighting that for us. I mean it's not something we want to get into a whole ton of detail on. But machine learning, deep neural nets, it's a very hot topic. People are used to seeing it in their lives of search engines and all manner of other things. But if you look at the range of things that we design algorithms for, that require -- that type of an approach to a single processing problem is real useful when it's -- you are trying to solve a real specific problem, you've got a real defined element of technical merit that you're trying to achieve on the other end -- on the other hand, there is lots and lots of things that fall into the category of where there are many, many used cases and really almost an unimaginable number of scenarios that can unfold. So for example, if you were using a noise canceling headset in an environment, you really -- there is no way to definitively list off all of the various sorts of noise that might happen in the environment where that headset is being used. And so designing kind of the use-case manager, what happens if a siren goes off and you are jogging along or what happens under this scenario or that scenario. It's difficult to approach from a traditional design-the-algorithm-type of approach and that's kind of one of the things that we've kind of undergone some learning as we've rolled out the ANC headsets. There are some categories of algorithm that you are better off with the machine learning approach where you characterize -- you just throw stimuli at it and you train on what is the desired outcome. In the case of noise canceling, for example, very simplistically, it could be no noise comes out of the headset. But in any event, there are categories of algorithms or categories of use cases that are better handled in a deep neural net-type of algorithm development environment. Whether that's the actual algorithm themselves or the tuning parameters, for example, for noise canceling or in this particular case, to where we want to start focusing the efforts with this recent acquisition on the voice biometric and specifically the performance in noise, that's the kinds of things. And this team we happened to find -- there is no shortage of people out there working on neural maps and machine learning, of course, but these folks -- it's relatively small team, it's not a huge acquisition. So we don't want to make too much of it. But these were some real sharp folks that worked well with our team. As usual, we approached it with a very bounded problem of, "Hey, here is a problem we are trying to solve. Why don't you see what you can do with it?" And then use the opportunity to work real well together -- to work together and see if the teams gelled and that happened. And we think they can help us with making the voice biometrics stuff work better and better in a noisy environment over time. So something we are excited about, and as Thurman mentioned, a great use of cash. Any time we can look at the existing strategy and find ways to accelerate it, pull it in a little bit, make it a little more certain, take some of the discovery out, that's by far the best way to -- we think that's by far the best way to use our cash. And we are happy to be able to find another one this past quarter.

Great. I appreciate the color. And then just thinking about wireless headsets, air pods have been a smashing success it looks like, so far. And I think we're going to see a lot more phones coming out over the next 12, 24 months with the new Bluetooth, where we can multi-cast. So I just wondered, is there a design funnel that's starting to come together for you guys in terms of anything you are going participate on the wireless headset side? Any update there would be helpful.

Yes. I mean we've been engaged, and we continue to be engaged in a number of Bluetooth opportunities. Our noise canceling device, for example, is interface agnostic. You can use it with any of the existing digital wired interfaces. The first product that's on the market is actually a 3.5-millimeter traditional headphone jack-type device, but you can also use the noise canceling with Bluetooth. I think as things stand today, if someone is trying to take a traditional Bluetooth radio and make a simple headset out of it, the opportunity there for us is not tremendous. But if you get down to the Bluetooth low energy generation, which is still kind of on its way, in terms of having full audio bandwidth ratified, then that really drives the bar a lot lower, in terms of power consumption we need to achieve a much lower level of power consumption and additionally offer some advantage in terms of fidelity over time that folks outside of our largest customer don't have advantage -- don't have the advantage of dealing with things in a captive kind of an environment. So we remain excited about headsets of all sorts, and I think over time, certainly, you are right, Bluetooth will play an increasing role in that over time.

(Operator Instructions) Your next question comes from the line of Rajvindra Gill from Needham & Company.

A question on the voice biometrics, Jason. I'm wondering if you're making a distinction within the artificial intelligence category for language between a voice recognition and natural language processing, where the voice recognition is primarily used in application, such as command management and connected homes, versus where natural language processing is the interaction between human speech and computer command, leveraging deep learning algorithm. So I'm just trying to get a little bit of a sense in terms of voice biometrics. How you are thinking about that opportunity since you are taping out the product this quarter, and obviously, view there is a big opportunity?

Sure. Yes. So the goal of our voice biometrics, just to be clear, is kind of neither of those things. I think coupled with a number of our other technologies, we can help make either natural language processing or command-type structures work better, if the device has certainty that the user is authorized. I think in a way that can be transparent to the user, we can make any of it. I think we have the opportunity to make those things work better. But just to be clear, what we are trying to do is identify the user not what they said. So for example, if somebody was to put our chip in a handset, you would go through an enrollment process, where you'd maybe say some -- say a paragraph with some particular phonetics in it and other things, but using free speech, meaning, without necessarily having a fixed phrase, so you could use it with any sort of natural language back-end that the user could simply say the "Okay, Google" phrase or whatever the unlock phrase happens to be, device wakes up, begins validating that you are authorized, and then go on to say something like, "Hey, read my e-mail that I just -- "I heard a tone, I got my e-mail." "Hey, read my e-mail." That's an application that's very, very simple, but I think most of us are really fixated on making sure whoever said, "Hey, read my e-mail" is actually me. If I lose my phone in a bar and somebody says, "Read my e-mail," that's a problem. I don't even know how many forms we have to fill out with the SEC if that happens. But it's probably not good. So we are not trying to figure out what the user said at all, and in fact, it's kind of the Holy Grail of voice biometrics to be able to validate the enrolled user without any knowledge of what it was they were saying, just knowing that the person is actually who was enrolled in the device. It's one of the most kind of fantastic things we've ever worked on. It's tremendously hard. It's not -- it's certainly the case that people have done this and worked on these kind of applications before. Some of the call center-type providers, service providers have this capability. It runs on servers. It uses racks of servers with stupendous amount of memory. And what we're doing is trying to take that down and render it into an embedded secure-type form factor. And we are very excited about the opportunities there. As I said, it's -- there is meaningful research left. It's -- on a lot of position, it is a slam dunk. But it is something that is very, very exciting, and something we can do a lot of different things with, once we get that out.

Definitely. And then just on the active noise cancellation road map, maybe you could talk a little bit about kind of where we are in terms of the road map as we get to the second half of this year and go into next year? The always-on listening chip that can differentiate between various noises without the need for any of these passive noise cancellations. What's been the feedback from the customers? And when do we think we are going to start to really move on the ANC headsets?

Yes. Well, the first device is out now. We expect to see more over the course of the year. I think we've learned a lot in the process of getting the first couple into production -- or the first one into production with more on its heels. We are trying to simplify the tuning tools. It's -- I mean, designing, in particular, the form factor you referred to is the nonsealed kind of ear bud, that's a pretty challenging feat, even using our devices for a bunch of acoustics people. So we're working on making that simpler. We are also working on rolling out versions for people who do want to use the -- say, a silicon kind of a grommet to seal up the ear canal, because there is demand for that form factor as well. We can simplify up some of the aspects of our device and work on a fixed filter-type performance. But again, it still achieves really stunning noise cancellation property. So simply -- basically, turning the crank on broadening the product line out, thinking about integration as it helps accelerate the product line as we go forward, we will keep working on that. But I see real good opportunities, and as usual, this stuff sometimes takes longer than it might appear (inaudible) on the outside, but it is easier to forget. I don't know if it's rocket science, but it's pretty close.

Your next question comes from Rick Schafer with Oppenheimer.

This is Josh Buchalter on behalf of Rick. Actually, just wanted to follow up on the previous question. Are you currently seeing most of your ANC engagements would you expect to be with handset OEMs? Or would they be more in the aftermarket?

We've seen a bit of both. We're trying to certainly focus on people that have accessory devices that they sell outside of -- coupled with the handset, but also could have the opportunity if they have a real strong customer response, they could think about them doing in-box. But we're trying to focus on the opportunities based on size and based on the customers that we think can really be good partners for us, or in some cases, already are good partners for us. So it's kind of varies, but as you can imagine, we are focused on things on order of size, which tends to couple with people that can -- that would have at least the option to in-box.

Got it. And then I was also hoping you could elaborate more on your longer term OpEx thoughts. Obviously, it's -- you've had some great years of 25% plus growth, but how should we think about OpEx trending in this year, which is expected to be a bit more modest? Should we -- is it more managing to your mid-20% operating margin target?

Yes, I mean -- yes, exactly, I had kind of alluded to that earlier. Relative to the top line growth, we will keep an eye on the R&D spending. Obviously, we want to continue to hire. And we think we've got some -- well, we know we've got some great opportunities to deploy those folks, but we'll certainly do that with the mid-20s operating margin in mind and the expected revenue growth in mind, and certainly, not ruling out the potential for some really good growth opportunities going forward. We definitely don't want to take our foot all the way off the pedal there.

(Operator Instructions) Your next question comes from the line of Tore Svanberg with Stifel.

First for Thurman. Thurman, revenue growth last year was strong, but I would say, cash flows were sort of spectacular, right? So was there anything in the fiscal year on the cash flow front that was unusual? And if not, what type of cash flow ratio to revenue growth should we think about going forward?

Well, there was nothing unusual in the year in terms of cash flow. I mean, it goes quarter-to-quarter, and it depends on -- you have things like paying down debt or doing share repurchases and all of that in terms of your cash balances. The cash flow from operations will continue to be strong through this year, and we don't expect any real change there.

And then, I would just add into that, that certainly, along with earnings, there is more leverage on that line. As we grow, we're trying to keep a pretty tight lid on SG&A, and we're a fairly conservative bunch, no matter how you look at it. So with earnings as well as cash flow generation, there should be some good leverage there as we grow. And we've certainly benefited a great deal from that over the past couple of years.

And then my last follow-up would be on the hi-fi deck. So it does sound like you're getting a lot of traction there. And I'm kind of surprised to see revenue is already now or at least production now after introducing that product not too long ago. So how should we think about that from a content perspective, per either a flagship phone or mid-tier phone?

I mean, I don't want to get too specific, because it's kind of a negotiation on a case-by-case basis. But I mean, it's something that, right out of the gate, it's something that's in the millions of units kind of right out of the gate, and we expect to do very well with that, especially in China over the course of the year. And frankly, we've got other versions where we can then take that device and market it into much higher performance line power gear. Because as far as I can tell, it's the best D/A Converter on the market and it consumes a tiny fraction of the power of what everybody else does. It doesn't need a whole basket of chip components, like some of the other solutions in this performance range would require. So I mean, it's a pretty compelling device, and the fact that some of the folks in China are still kind of trying to differentiate in a meaningful way on audio performance makes it a great opportunity in that market. So we're excited about it. I would say, we were excited about it before we got the device back, and the engineering team did a remarkable job of getting a production-worthy version out on early silicon, and it's been even better received than we might have anticipated. As we have always said, the kind of the stickiest sockets for us both on the -- both, which is a negative one you are trying to win and it's a huge positive once you are in there, smart codecs are pretty, pretty sticky. It's a big design effort and very intertwined with drivers and the operating system and everything else. Hi-fi, D/A Converters, when you are trying to win, turn into revenue, is much more of a -- it's a little bit more self-contained and not quite as intertwined with a whole lot of other system design. And I think there is very few other companies out there that will be able to duplicate the kind of performance and power tradeoffs that we have achieved with that device. So we're excited about it. Still kind of early innings, but it got out of the gate strong as you noted.

There are no further questions at this time. I will now turn the call back over to Chelsea Heffernan.

Thank you, operator. There are no additional questions. So I will now turn the call back to Jason.

Thank you, Chelsea. In summary, we are extremely pleased with our performance in FY '17 as we delivered substantial revenue growth, introduced a range of innovative new products and expanded our customer base. With the road maps spanning the complete audio signal chain, Cirrus Logic expects to leverage our solid customer relationships and engineering and execution prowess to drive long-term shareholder value. I would also like to note that we will be presenting at the Stifel Technology, Internet and Media Conference in San Francisco on June 6 at 9:45 a.m. Pacific Time. A live webcast of this event will be available at investor.cirrus.com. If you have any questions that were not addressed, you can submit them to us via the Ask the CEO section of our Investor website.

I would like to thank you, everyone, for participating today. Goodbye.

This concludes today's conference call. You may now disconnect.