InvenSense (NYSE:INVN) has just posted fourth-quarter results, and shares are down about 20% after-hours. Although I've been long-term bullish on the company, I had voiced caution about the share pricing going into this earnings report. The market now seems to be overlooking the fact the company beat consensus revenue projections and that two out of the three reasons for the EPS miss are actually quite positive. We saw a similar situation two quarters ago, when I had also warned of overbought conditions going into earnings, but went on to map out the long-term growth plan. This article will examine the factors that contributed to the EPS miss, and look at the next steps which the company is already taking to continue innovation and growth in the absence of the buyout scenario discussed in my prior article.
What Just Happened?
As mentioned above, InvenSense actually grew revenue almost 7% year-over-year, and the company also increased market share dramatically. So why the earnings miss? There seem to be three main factors:
- a larger % of legacy product in the sales mix, this is expected to decline in the coming quarters, thereby returning margins to their usual levels.
- "invested considerably" in new customer qualification activities, which are expected "to yield significant growth and market share gains" in the coming year.
- increased R&D expense.
The "new customer", referred to in point #2, is most likely Apple (NASDAQ:AAPL), whose extensive qualification processes are well known. I've already written quite a bit on why earlier analyst reports of InvenSense getting that design win were wrong and why it will only happen when Apple introduces a wearable device, which is currently expected in September, but could be confirmed at Apple's WWDC as early as June 2. Virtually the only other candidate for such a device would be QuickLogic's (NASDAQ:QUIK) sensor hubs, and InvenSense offers cheaper, lower-power solutions which are much better suited to wearables. You can see detailed discussion contrasting these two very different approaches in the comments of my prior article. For the purposes of this one, suffice it to say that I maintain my prediction that InvenSense will gain Apple as a customer this year, and that one reason for the miss on the bottom line actually supports that claim. It's also very important to note that in the conference call, InvenSense management characterized guidance as very conservative and specifically said that it does not include any new customers, so when this prediction does pan out, it should be a major catalyst for INVN.
What's About To Happen
Instead of rehashing Apple speculations, though, I'd like to focus on point #3 and my next prediction for InvenSense. When I first wrote about the company over a year ago, I talked about how mobile devices are transforming our lives and how they would gradually transition to wearable devices. What we've actually seen in the intervening year, though, is that mobile devices are bifurcating. On one hand, mobile devices are actually getting bigger to accommodate larger batteries and screens (a point my next article will address). On the other hand, we see the gradual evolution of wearables, with a host of new devices introduced at the Mobile World Congress in March. Still, we've yet to see the killer application that results in their adoption across most of the rich world, the way we did with mobile and smartphones. For that, these two classes of devices need to start working together.
Believe it or not, I don't own a smartphone; I use other portable devices for information management on the go. I even went so far as to buy the "dumb" phone that I use now several years in advance, and kept it unused until its duplicate broke or got lost, just so I would have a good replacement ready. The problem is that smart"phones" are invariably awful when one tries to use them as actual phones, largely because the microphone is far from your mouth, and/or the speaker is far from your ear. As a result, call quality is terrible, because both ends of the conversation pick up too much ambient noise, and many people communicate by text or email, even when the situation makes that very inefficient. Furthermore, the bigger the smart"phones" get, the worse the call quality gets. I know I'm not alone here, I've heard this complaint from many others.
When I predicted that InvenSense would transition from sensor provider to SoC (system-on-chip) designer, at the end of last year, it was in part due to the ADI microphone acquisition. In the Q4 conference call last night, management described audio processing and contextual awareness as "highly complementary". Think about how many people you already see walking around wearing ear-buds, or other speakers. Now think about one of those microphones in a smart watch and how that completely solves the problems of call quality. No more reaching; it's easy to get the positioning right, and your hand and wrist create a natural shield to block external sound and reflect your own speech into the microphone. So, I did some checking, and lo and behold, InvenSense is already on it, with a job posted months ago. This goes far beyond phone calls, too. The configuration I'm speaking of makes all of speech recognition easier and applicable in more situations. InvenSense is about to transform how we interface with our devices by integrating microphones and simple speech recognition into their always-on, contextually-aware SoCs. This change in interface will be so powerful that I might just have to junk my current phone before it breaks.
Where We're Going
This transformation isn't going to happen overnight, but markets are forward-looking; I think the products being introduced later this year will start to make it very clear that this is the direction that the next jump in mobile innovation will take. In fact, both Apple, with the M7 co-processor in the latest iPhones, and Google (GOOG, GOOGL), with the Moto-X, have begun to move in this direction; now it's integration of these features with a wearable that's the missing link. The first InvenSense SoCs will probably just recognize simple motions and commands to know when you want to wake up the more power-hungry processor in your mobile "phone" and get it to do more complicated things. Actual speech recognition is very computationally-intensive, and would certainly require other components. For InvenSense to get there, it would require acquisition by a large company, like Intel (NASDAQ:INTC) or Qualcomm (NASDAQ:QCOM), or merger with another specialist like Cadence Systems (NASDAQ:CDNS).
Make no mistake, though, more and more functionality will continue to move onto SoCs due to their power efficiency. This will continue to be the trend until conductive, generating fibers start to become the norm in clothing, and that is likely beyond the investment horizons of any but those investing for their far-off retirement or children. InvenSense is by far the best-positioned company for providing the always-on, contextually-aware functionality that this interface shift will require, but it will take time and lots of work. Until we can lose those bricks in our pockets completely, sensor hubs like those mentioned from QuickLogic will probably play a complementary role, since their strength is in customization and flexibility.
In the meantime, InvenSense has plenty of other growth factors that indicate to me that the after-hours sell-off is overdone:
- Management indicated that OIS (optical image stabilization) adoption is still at the beginning stages. The trends towards conferencing and shooting video with mobile devices will drive OIS towards being a must-have feature.
- The newest InvenSense products add pressure sensors, which aid in navigation, and will move InvenSense further up the food chain.
- Demand from lower-end Chinese producers is expected to rebound, with Xiaomi re-emerging as a 10%+ customer.
- The latest versions of Android specify that real-time sensor monitoring must be offloaded from the main processor, which is exactly what the SoCs from InvenSense do. Expect continued market share gains across the Android smartphone ecosystem.
All of these factors have allowed management to guide for 25-35% revenue growth over the next year, which excludes new opportunities and is characterized as conservative. Furthermore, having 30% of float shorted, with a rebate rate of 7%, is sure to mitigate the downside and aid a short-term rebound in share price. In fact, if the stock does not rebound sharply in the near term, it will become a prime candidate for a short squeeze later in the year.
Over the longer term, the introduction of Android Wear by Google will greatly accelerate wearable device proliferation. Since no current vendors can effectively compete with InvenSense for this form factor, this is where earnings should really take off. The company's patented Nasiri fabrication process represents a formidable moat to new competition, as well. After looking back at my previous articles and seeing how management has deftly eliminated each of the risk factors I listed, I maintain my 1-year $24 price target for INVN. Other analysts agree; Needham has just upgraded from hold to buy, and Goldman has reiterated their buy rating. This may be a very short-lived dip.
Disclosure: I am long INVN. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article. I may initiate a long position in CDNS and/or QUIK in the next 72 hours.