Last Thursday after the close, Anadigics—a supplier of RF components to the handset and cable set-top box markets—cut its forecast for third quarter revenue from $78 million to $64 million, 18 percent lower than the guidance management established a mere three weeks prior (see Q2 earnings call transcript).
Predictably, ANAD shares were hammered Friday as the stock lost 38 percent of its value. Equally as predictable, sell-side analysts limited their analysis of the situation at Anadigics to the near-term implications of the operating issues brought to their attention by management while, in our opinion, ignoring what we believe are larger existential threats to Anadigics that are in the process of surfacing.
Management has attributed the sudden and drastic change to its near-term outlook to the realization of an inventory build up at its major customers. The inventory build-up, so the narrative goes, stems from a prior shortage of parts that caused Anadigics to place its customers on allocation. This parts shortage in turn was created by the Company’s inability to bring up needed manufacturing capacity at its Warren, NJ facility in late-2007. Anyone familiar with the semiconductor industry knows that when customers are placed on allocation they typically inflate their orders so that they can get the parts required to satisfy their current demand while also building inventory so they can manage through the shortage and avoid their own supply disruptions. More consequently to the offending supplier, customers also immediately turn to secondary suppliers for both existing designs and for future designs. A “fool me once, shame on… ain’t gonna get fooled again” dynamic.
So, Anadigics’ management paradoxically claims that during the period of a parts shortage it shipped too much and now that excess inventory needs to be burned off. Management also confirmed what had been previously learned, that the Company has lost important reference designs with Qualcomm (QCOM), with whom, prior to the aforementioned operational issues, it had previously enjoyed a “most favored” partner relationship. To cap the bad news, management indicated that due to its damaged competitive positioning it will forgo further investment in its new China facility beyond what is already committed.
At $3.69 per share and approximately half of its market cap ($1.80 per share) represented by its net cash, the stock appears cheap for a company undergoing a short-term inventory issue and momentary loss of market share. For most sell-side analysts this is where the story ends and on what the investment conclusion pivots. And if a short-term operational bump is all there was to the Anadigics story, ANAD shares at 80 percent off their 52-week high might be tempting. But in our opinion that is not all there is to the story. In fact, it is not the real story at all.
Long-Term Existential Issues
As noted, Anadigics sells to three primary end markets: it sells front end components, namely power amps and RF switches to wireless handset makers; it sells RF tuners and splitters to cable set-top box makers; and it sells WiFi products for 802.11 standards. We think the Company faces longer-term existential threats in the first two of these three markets caused by technological changes outside of its control.
First, let’s look at the wireless handset market. The technological guts of a wireless handset can be divided into three separate modules: 1) the digital baseband, which is dominated by Qualcomm, Broadcom (BRCM), and Texas Instruments (TXN), 2) the transceiver, where Silicon Labs (SLAB), Skyworks (SWKS), Analog Devices (ADI), STMicro (STM), and others compete 3) and the RF front-end where Anadigics competes against RF Micro Devices (RFMD), Skyworks, Triquint (TQNT), and others.
There is an extremely important difference between the first two modules—digital base band and transceiver modules—and the front-end RF module: the former are manufactured in standard CMOS process while the front-end RF module is not. The practical importance of this difference is that leading CMOS processes provide for mind boggling transistor densities, which enables the digital base band and transceiver modules to be fully integrated into a single chip. However, the demanding thermal requirements of the RF front end module have not been accommodated by traditional CMOS bulk substrates. As a result, while the digital base band and transceiver modules have been integrated into a single ship on CMOS, the front end module still consists of numerous discrete parts based on a variety of class III-V (non-CMOS) compounds—GaA power amplifiers, pHEMT-based antenna switches, ceramic duplexers, isolators etc. This is the market that Anadigics serves.
Given the obvious advantages that a fully integrated chip has over a solution based on discrete parts—lower cost, reduced size, and less power consumption—it is only logical that a substantial effort has been made to develop this disruptive technology for the RF front end.
CMOS is coming to the front-end RF module. Peregrine Semiconductor—a private company we have been following for quite a while now—has developed integrated RFICs based on standard CMOS which promises to disrupt the current RF front-end. While the thermal limitations of bulk substrates has created an opportunity for suppliers of discrete RF parts based on esoteric class III-V compounds, Peregrine has perfected the use of sapphire as a substrate material which is capable of accommodating the electrical and thermal demands of the RF front-end module. A fully integrated front-end RF module will leave Anadigics and other providers of discrete front end parts out in the cold.
For an appropriate parallel of this dynamic consider that SiRF Technology, a provider of GPS chipsets for personal navigation devices and GSM handsets, never sold GPS chips to CDMA handset makers because the GPS function it provided was already an integrated feature of Qualcomm’s baseband chip. Peregrine poses a similar threat to Anadigics (and other providers of discrete RF parts) as an integrated RFIC will bring all the functions of the RF module to the handset manufacturer at a dramatic cost and performance benefit thereby leaving no market for discrete RF parts.
While Peregrine is a private company without public financial disclosures, we believe the company is on track to generate $85-90 million of revenue this year with its unit shipments more than doubling in each of the past three years. This would suggest that the Company, on a revenue basis, is already more than one-third the size of Anadigics and growing at over 50 percent annually. While we don’t know if Peregrine is a direct source of Anadigics current market share losses, we do know that Peregrine’s rapid success is certainly coming out of someone’s front-end hide. Regardless of the current impact Peregrine might be having on Anadigics, the longer-term ascendance of Peregrine does not bode well at all for the long-term fundamental health of Anadigics.
IP Protocol is coming to cable set-top boxes. Anadigics second end market is the cable set top box where it sells RF tuners and splitters to the cable set top box manufacturers. Over the past few years Anadigics has been rewarded handsomely as the cable industry transitioned from analog to digital networks. Digital set top boxes require two tuners instead of just one in the legacy analog boxes which has essentially doubled Anadigics revenue content per set top box. There is however another technology transition rapidly unfolding in cable; this one promises to altogether eliminate Anadigics’ revenue opportunity in the set top box market: the transition from RF to IP based networks.
IP based, or packet networks, enable carriers (telecom and cable) to expand the capacity of their networks far beyond what is practical over current RF networks. There is a current rush to expand network capacity driven by video streaming, the demand for more high definition programming, and by the competitive pressures that telecom companies are beginning to exert as they roll out their own feature-rich IPTV services.
The implication for Anadigics of the transition from RF to IP protocol in video networks is quite simple: IPTV set top boxes don’t require RF tuners, therefore they don’t require Anadigics. The faster cable networks transition to IP architecture the faster this market will disappear for Anadigics, but regardless of how fast it will disappear, it will disappear.
Investors evaluating the merits of an investment in Anadigics after Friday’s drubbing are best to consider that Anadigics problems are not simply short-term bumps—mere inventory problems and market share losses that are quickly righted in a quarter or two—but rather they should be considered early manifestations of very real and ultimately inevitable long-term existential challenges for Anadigics.
Disclosure: Author has no position in Anadigics.