Microchip (MCHP) makes microcontrollers and related products. When I last wrote about MCHP [See Microchip Ups Guidance, June 4, 2013] I said "on the customer side, inventories are believed to be extremely lean, which could result in increased buying if final demand shows signs of improving with the global economy."
Second quarter results broadly confirmed that story. Revenues hit another record at $462.8 million, up 8% sequentially from $430.1 million and up 31% from $352.1 million in the year-earlier quarter. GAAP net income was $78.6 million, up 32% sequentially from $59.7 million, but nearly flat from $78.7 million year-earlier. GAAP EPS (diluted earnings per share) were $0.37, up 32% sequentially from $0.28, but down 5% from $0.39 year-earlier.
Microchip closed on July 7th at $41.01, near its 52-week high of $41.78 on August 1. The 52-week low was $28.92 on November 16, 2012. At $41.00 the $0.354 dividend per quarter works out to 3.43% annual yield. The ex-dividend date is coming up on August 19. Microchip frequently raises the dividend by a fraction of a penny.
The story behind record revenues? There was the successful acquisition of SMSC last year, but most of the growth from that was in by the end of 2012. The simple story is increasing demand for Microchip microcontrollers. This in turn results from a loyal base of engineers, a base in legacy applications, and a very wide variety of models that meet a broad breadth of engineering needs. However, Microchip has a relatively small presence in the PC, tablet and smartphone spaces. The PC space has been suffering a downturn since 2012. Microchip microcontrollers are used in industrial and medical applications. The reviving housing market may be helping since Microchip has a strong presence in household appliances. It also sells into the automotive industry, which has been ramping. New analog chips and chips combining analog circuits with microcontrollers have been driving new design wins.
Some sell side (Wall Street) analysts have questioned the viability of Microchip on two fronts. Before 2012 the main line of attack was based on data-width. As with CPUs, microcontrollers typically process data in 8, 16, or 32-bit chunks (CPUs now mostly are up to 64-bit chunks). Successful in the 8-bit and then 16-bit markets, Microchip was slow to enter the 32-bit arena. At the time they said they would make more 32-bit chip variations when the demand was in place. In the June quarter 32-bit microcontroller revenue was up 26.3% sequentially to a record, and up 362% y/y. So entering that space carefully has not held them back. Microchip's 8-bit and 16-bit chips continue to generate record or near-record revenue each year.
The other objection has been that many competitors have shifted to ARM based cores, the same cores used in smartphones. But ARM-based cores are not specifically designed for microcontrollers. Results show engineers prefer Microchip's PIC cores for many applications. Which is better is an application-specific question. ARM devices can run Linux, including Android, but microcontrollers often are programmed from the ground up for efficiency (to minimize memory requirements); they run better without an operating system.
If there is a problem with the September quarter, it may be in meeting demand. Microchip's own foundries are back to full time operation. Bookings at the end of the quarter were high. Lead times (from ordering to shipping) are lengthening; management characterized them as less than optimal. Microchip now pays other foundries to make some of its chips, so there is some flexibility there, but less control over timing. Guidance for September quarter revenue is $472.0 to $490.6 million, which would be a record even at the low end.
Given all that, is MCHP priced reasonably for long-term investors, at $41.00 per share?
GAAP EPS for the last 4 reported quarters totals $0.62. At $41.00, that gives a trailing P/E of 66.1, which seems high even given the level of growth. Non-GAAP trailing EPS is $1.98, which works out to a trailing non-GAAP EPS of 20.7, which seems low for the level of growth expected. Which one to use? In a typical quarter non-GAAP eliminates non-cash accounting expenses, notably share-based compensation expense and acquisition-related amortization, which is fair enough as long as you understand that. The big divergence between Microchip's trailing GAAP and non-GAAP came in the 3rd and 4th calendar quarters of 2012. In addition to the regular excluded expenses, substantial non-amortization acquisition related expenses were excluded. These were from the SMSC acquisition. While such expenses are real, they are not predictors for future run rates. For predicting future profits, the non-GAAP trailing EPS of 20.7 is a better indicator, as long as we keep in mind the dilutive nature of stock based compensation.
Guidance for calendar Q3 is to GAAP earnings of $0.40 to $0.44. At the midrange, $0.42, P/E would be 24.4. Non-GAAP estimate is $0.58 to $0.62, with a midrange of $0.60 leading to a P/E of 17.1. Considering the size of the dividend and growth prospects, to me that would indicate $41 per share is at the very low end of the current reasonable range for Microchip.
If you want a solid semiconductor stock that pays dividends and grows revenue, earnings and dividends on a fairly consistent basis, MCHP should be on your short list for further research. I have held MCHP since 2007, when I first bought it at $32.70 per share. While it has not been near my best percentage gainers, I like the relative lack of volatility (for a tech stock) and the reliability of the dividend.
When researching a company, it is a good idea to research the competition. Some other large microcontroller manufacturers trading in the U.S. are: Atmel (ATML); Texas Instruments (TXN); Freescale (FSL); and STMicroelectronics (STM), Silicon Laboratories (SLAB). Renesas is a major competitor that trades on the Japanese market.