Qualcomm (QCOM) turned in a solid quarter on April 24 with total GAAP revenue of $6.124 billion and operating income of $1.877 billion, both up 24% year over year. The market rewarded Qualcomm with a 5.4% drop in price the following day on heavy trading. The spike in volume at the opening suggested buy side fund managers had dumped the stock on the assumption that Qualcomm's price was as good as it's likely to get for a while. They're right, but for reasons that go beyond what have been described in the media.
The media have tended to focus on elements that could be interpreted as "disappointing" in the earnings report, under the assumption that investors must be reacting to something in the report. The two supposedly disappointing aspects frequently pointed to were the full year earnings guidance and the lack of earnings "upside."
Qualcomm actually increased its full year revenue guidance to $24-25 billion, up from $23.4-24.4 billion and its GAAP EPS to $3.78-3.93 from $3.61-3.81. Earnings aren't expected to grow quite as fast as revenue, 23% vs. 31% at the high end of the guidance range, indicating some squeezing of margins.
The slowdown in earnings growth is particularly noticeable in the chip production group, Qualcomm CDMA Technologies (QCT), where quarterly revenue of $3.916 billion grew 28% year over year, while earnings before taxes of $681 million only grew 14%.
A longer-term problem
These issues are indicative of a longer-term problem that Qualcomm has as a commodity producer of ARM-based systems on chip (SOCs) that are used in at least some models of every major brand of smartphone, with the exception of Apple (AAPL). This problem is illustrated in IDC's data on smartphone market share released the day after Qualcomm's earnings report. In the table below I excerpt data from IDC's Q1 2013 and Q1 2012 reports for the top five vendors:
Top 5 Vendors 2013 Q1
2013 Q1 Unit Shipments
2013 Q1 Market Share
Top 5 Vendors 2012 Q1
2012 Q1 Unit Shipments
2012 Q1 Market Share
Research In Motion
Nokia, Blackberry, and HTC use Qualcomm SOCs in their latest products, but the fact that all three have dropped out of the top five rankings wouldn't necessarily be bad for Qualcomm if the new members of the top five also made as extensive use of Qualcomm SOCs. LG and ZTE do in fact use Qualcomm almost exclusively, but Huawei is mixed, with some products using Qualcomm and some using processors designed and made by Huawei itself. This is the trend that really threatens Qualcomm.
This trend is made even more threatening by the growth of Samsung, which is well known as an ARM processor manufacturer that also provides foundry services for Apple for its current generation of iOS devices. Samsung's product portfolio is currently mixed, with some phones having Qualcomm SOCs, and some having Samsung's own chips. The current Galaxy S4 is split between a Qualcomm quad core SOC for the U.S. market and a faster, more advanced eight core Samsung processor for the rest of the world.
The oft repeated mantra
Although many tech pundits repeat the mantra that "mobile devices will become commoditized" under the assumption that "custom" SOCs such as Apple's are a short-lived aberration, in fact, that isn't the trend. Three of the top five smartphone vendors produced their own silicon in 2013 Q1, as opposed to two out of five the previous year. This trend will only accelerate, for the following reasons:
1) The growth of foundry services allows even small chip design houses to realize many of the economies of scale of large chip producers, because foundries can spread capital equipment costs over many customers. Even Qualcomm uses foundries to produce its chips.
2) The commoditization of design IP allows smaller companies to be SOC integrators. Companies such as Apple don't have to design a processor from scratch but instead license the designs they need from others such as ARM Holdings (ARMH). Thus the systems integration role of device makers has moved to the level of the chip, affording customization of the SOC to meet the design goals of their devices. Apple's A-series SOCs that power all its iOS devices have been derided as merely catering to Steve Job's penchant for "boutique processors." In fact, they merely represented the logical extension of Apple's role as a computing device systems integrator.
3) Customization of SOCs affords a way to optimize mobile devices and provide product differentiation. Device makers such as Apple and Samsung can provide an optimal performance/power consumption trade that maximizes battery life. SOC customization allows devices to be differentiated in terms of hardware interfaces and features.
As processor design expertise has become more widespread, as processors integrated more functionality, as foundry services became commonplace, it was inevitable that the lock on processor manufacturing that commodity producers such as Intel (INTC) had would be broken. Qualcomm has been a commodity producer of ARM SOCs in the Intel vein, and this has worked out for it as the industry transitioned from the commodity CPU model to the custom SOC model. I doubt this will work much longer. The exit of Texas Instruments from the commodity ARM processor market last November was the handwriting on the wall.
Why haven't custom SOCs completely taken over the mobile business? Because the smart device industry has been growing so fast that Samsung and Apple couldn't keep up with demand. As the market matures and demand slows, Samsung and other SOC integrators will increasingly prefer their own silicon.
To be sure, Qualcomm doesn't just make ARM SOCs, but also has a thriving business in CDMA and 4G LTE radio technologies. Many phones that don't use Qualcomm SOCs nevertheless use Qualcomm radio devices, such as the iPhones. Qualcomm lumps revenues from its radio chips and SOCs together under the QCT division, so it's impossible to say with certainty what the revenue breakdown is, but I would guess that radio devices are about a third of the QCT revenue and earnings.
Qualcomm also has a thriving technology licensing business, which had $2.057 billion in revenue and $1.803 billion in earnings before taxes. The revenue streams from wireless devices and licensing are certainly enough to sustain the company, but I expect gradual erosion in SOC revenue starting in the third or fourth quarters. The accompanying slowdown in growth of the QCT division revenues will come as a rude awakening for Qualcomm's investors by Q4. But even then, widespread assumptions of the dominance of the commodity processor model will keep most investors' doubts at bay, so that the post earnings report drop will only be in the 5-10% range.
The worst is yet to come in 2014 as QCT revenues actually contract. Given that Qualcomm's valuation is predicated on rapid growth, a share price collapse next year appears inevitable.