InvenSense, Inc. is a pioneer and global leader in devices and related software for the motion interface market that detect and track and object's motion in 3D space such as gyroscopes, accelerometers, compasses, and pressure sensors, as well as microphones through the acquisition of the MEMS microphone product line of Analog Devices. The company's products have broad applicability and target consumer electronics applications such as smartphones, tablets, ultrabook and laptop computers console and portable video gaming devices, digital still and video cameras, smart TVs (including digital set-top boxes, televisions and multi-media HDDs), navigation devices, industrial sensors, toys, and health and fitness accessories. Most of it's revenue (79%) comes now from mobile devices (it was 56% in 2011), and it's biggest customer is Samsung (35% of revenues).
The company has a patented fabrication platform which impacts the cost of product packaging and testing, which reduces the cost of the end product. They claim this provides a significant competitive advantage. Invensense has 194 R&D employees and spent 48$ million on research last fiscal year. It's main competitor is STMicroelectronics, which has $8 Billion in sales, spends 1.8$ Billion on R&D, has 9000 employees in this department and holds over 16 000 patents. The customers of STMicroelectronics include Apple, Blackberry, Nokia and also Samsung. Other competitors are Analog Devices, Inc., Epson Toyocom Corporation, Kionix, Inc. (a wholly owned subsidiary of Rohm Co., Ltd.), MEMSIC, Inc., Murata Manufacturing Co., Ltd., Panasonic Corporation, Robert Bosch GmbH, Sensor Dynamics, Inc. (acquired by Maxim Integrated Products, Inc.) and Sony Corporation.
If we take a look at the past 5 years, revenues have skyrocketed to 252$ million, with mobile being 79%, optical image stabilization 14% (8% in 2012) and gaming 7% (36% in 2012). The company has clearly caught the trend in mobile devices, while it's gaming business totally deteriorated (from 55$ million in 2012 to 17$ million now).
As the company acknowledges in it's 10-K: "The average selling prices of our products continues to decrease, which could have a material adverse effect on our net revenue and gross margins if we cannot reduce our costs. Our primary customers expect the average selling prices of our products to continue to decrease over time as a result of agreements we enter into with our customers from time to time, competitive pricing pressures, new product introductions by our competitors, shifts in customers’ product architectures, product end-of-life programs and for other reasons.
2. Rising inventory
Last year, inventories increased from 23 million, to 73$ million, a 213% increase. In the last conference call, the CEO Behrooz Abdi explained, that this should help them "prepare for multiple significant customer ramps in the coming quarters". Which ramps do they mean exactly? Revenues grew 21% for the full fiscal 2014, 13% and 7% for the last 2 quarters. For the next quarter, the company forecasts revenue of 63-66$ below consensus and previous estimates of 69$ million. For the full year, Invensense expects revenue growth of 25-35%, the slowest in four years. DIO (Days inventory outstanding), a measure that tells how long a company keeps inventory before they sell it, has increased from 18 days to 104 days now.
So here we have a company that builds inventory for "future growth prospects" and at the same time decreases revenue estimates and expects the slowest growth in years? Semiconductor inventory becomes obsolete very quickly, so they are either expecting huge orders in the next few quarters (which they are not) or they are manufacturing products that no one is buying (which I think is the case). This inventory will have to be written down in the future, depressing profits.
3. Declining profits and cash flows
While revenues have shot up in the past, operating cash flow has declined from 44$ million in 2012 to negative 11$ million in 2014. Because they don't generate much cash, Invensense has borrowed 175$ million last November, to fund "working capital, capital expenditures and general corporate purposes". Interest costs from this debt will decrease profits further. The company is spending increasing amounts on R&D (48$ million last year, an increase of almost 100% percent) which might be either a positive thing for the future, or a result of fighting tough competition. The semiconductor and MEMS industries are characterized by companies that hold large numbers of patents and other intellectual property rights and that vigorously pursue, protect and enforce intellectual property rights. Last year INVN paid 15$ million in claims for infringing a STM patent.
INVN is also spending more and more on marketing costs, which were 10% of sales in 2010, but now represent almost 20%. Again this comes as the revenue growth is slowing so I think it's a result of a really competitive environment where it's hard to acquire customers. The company has missed EPS estimates in the last quarters (0.15 vs 0.17, and 0.07 vs. 0.1) and downgraded expectations for the next quarter to 0.07-0.08, while analysts were expecting 16 cents.
4. Valuation
INVN trades at a P/E of 66, P/S of 6.1 and P/B of 5. Analysts expect an EPS of 0.7 USD next year, which would imply a 26 Forward P/E. The consensus target price is 22.5. Wall Street also expects a 22% 5-year EPS growth rate, and a decline in sales growth to 19% in 2016.
As explained above, I believe earnings will be well below estimates, my belief is around 40 cents per share for 2015 (and lower in the following years), implying a forward P/E of 46. I also think it's a highly cyclical business, their earnings (or losses) and sales will fluctuate wildly in the coming years and as such it's very hard to establish a valuation. The closest competitor, STMicroelectronics trades at a P/S multiple of 1.1, but it's a large company with anemic growth (5% +/-) prospects. INVN future growth is a bit better, but I still don't expect more than 15% in the next 5-10 years. There are rumors that iPhone 6 or iWatch may come with INVN sensors, but so far nothing confirms that.
So the question is, how much would you pay for a cyclical business in a very competitive field, with a 15% expected growth rate and erratic profits? We can assume INVN will grow 3 to 4 times faster than STM, using a P/S of 3.3 and 4.4 we would get a target price of 9.3 USD to 12.3 USD which is a level it traded at just a year ago. As I always wait for stocks to go down before I short, I will initiate a short position in INVN if it declines below 17 USD.