This column was originally published on RealMoney on Aug. 2 at 4:17 p.m. EDT. It's being republished as a bonus for TheStreet.com readers.To read Part 2, which was also republished as a bonus, click here.
In the week or so I've been writing for
, some emailers have wondered why I dislike
. It's an excellent company, but from my perspective, it's faced with some very serious problems that investors have ignored.
I've summarized the issues that I believe are critical to the company here, and in a column that will appear Thursday.
In short, Qualcomm:
- Lags its competitors from a technology perspective, a deficiency that hinders its ability to offer competitively priced parts without hurting gross margin;
- Does not compete effectively in the highest-growth markets; and
- Remains overvalued as a pure royalty play.
What is technology leadership? This question seems to draw the most fire because readers are always pointing to Qualcomm's cutting-edge HSDPA (high-speed downlink packet access, or 3.5G) and HSUPA (high-speed uplink packet access, or 3.75G) products.
Although those ultimately will become markets within a few years, today those services are largely in their infancy. When they represent a material percentage of the wireless industry, and if Qualcomm has the No. 1 position in these parts, I'll be happy to recognize the company's technological leadership. But Qualcomm is a technology laggard in a number of other areas.
Processing Process Technology
Process technology plays a critical role in the cost of parts and the price of handsets. The largest part of the bill of materials (BoM) in the handset is the silicon, and the cost of parts is a function of the area of silicon used. Consequently, companies that manufacture on leading-edge processes have a distinct cost advantage over those on lagging technologies, assuming the same basic functionality.
That advantage can manifest in materially lower costs to increase market share, higher margins or some combination of the two. The table below shows for demonstration purposes the magnitude of the advantage.
A company that was sitting comfortably with 60% gross margin on a 130 nanometre part is suddenly faced with a competitor that has a functionally equivalent part on the next-generation process that can be priced significantly lower. Of equal importance is the severe hit to gross margin if Competitor A matches pricing in order to avoid lost market share.
Qualcomm was fully two years behind the field in bringing its silicon to 90nm and a full year behind on 65nm. Originally, the size difference mattered little among its CDMA carriers within developed markets where Qualcomm had 90%-plus market share of chips.
However, as the company ventured into the WCDMA segment, there were numerous competitors with at least one chip (such as
) that was substantially farther along the process technology curve than Qualcomm.
The issue has also reared its head as GSM and CDMA carriers battle for subscribers in emerging-growth markets, which I'll discuss in greater depth in Part 2 of this column.
The company has accelerated its efforts in recent months, but is still playing catch-up, in part because its foundry partners have a variety of different objectives to address.
The bottom line is that Qualcomm-based handsets, whether CDMA or WCDMA, cost more than their functional equivalent based on silicon from other suppliers.
I believe that with the advent of WCDMA, Qualcomm anticipated worldwide market share would begin to fragment with the top five (
) losing out to the second-tier original equipment manufacturers, particularly within developing markets like China. This would fit quite well with Qualcomm's merchant market strategy of selling standardized silicon to all.
Unfortunately, the exact opposite has happened thus far; market share is increasing among the top tier at the expense of the lesser players. In order to achieve its publicly stated objective of 50% of the WCDMA silicon market, Qualcomm will have to crack Nokia, Motorola or EMP in order to have any chance.
However, because each requires a certain degree of customization as well as having already made a very large investment in software for current chipsets, the task will be difficult.
At last year's JP Morgan conference, Qualcomm's then-president-elect Steve Altman did not dismiss the notion of offering Motorola some type of "sweetener" to entice it to use Qualcomm's chips for WCDMA. Since then, Motorola has publicly called Qualcomm solutions "too expensive" from a number of podiums.
Given the fact that Motorola already sources silicon from
and Texas Instruments, I guess we know who's in the driver's seat. To me, this is a classic example of "be careful what you wish for." Given Qualcomm's high-cost parts and Motorola's hardball strategy, any "success" may be a public-relations plus and a financial drain.
Apparently unknown to or ignored by the investment community has been the integration side of technology leadership. Silicon integration is the key to cost reduction for handsets, not just in emerging markets but also around the world. It reduces total silicon cost, reduces manufacturing costs, increases performance characteristics and increases reliability.
I have frequently used the history of Ethernet (EN) LANs as an analogy. Much like handsets, EN cards started with hundreds of parts and cost many hundreds of dollars. But each company leapfrogged the competition with increasing levels of integration, driving costs down until we finally arrived at a single-chip solution. What we as users received through relentless integration is a service that operates at more than 300 times its original performance specs at 1/25th of the price.
Translating this experience into the handset world we see
and Texas Instruments all with single-chip products in various forms. To be fair, we're not talking about one chip, a screen, some buttons and a speaker; other components are required. But the bulk of the electronics BoM has been integrated into a single piece of silicon.
Qualcomm has also talked about its "single-chip solution" for more than a year. However, prior to announcing its QSC series (Qualcomm Single Chip) of parts, management acknowledged that this is a SiP (system-in-package): three individual pieces of silicon contained in one package. It will cost more, not scale, be larger and have lower performance characteristics than its single-chip competitors.
One of the primary reasons Qualcomm is not able to deliver a true single-chip solution is the lack of digital RF (radio frequency) technology, which the others have spent quite a few years developing. Is Qualcomm close?
Unfortunately, it appears that Qualcomm isn't even in the ballpark on digital RF. A recent article published in
Mobile Handset DesignLine
by engineers at the highly regarded semiconductor analysis firm Semiconductor Insights presented the results of their search of the U.S. Patent Office for digital RF-relevant patents. Texas Instruments, along with a half-dozen others, were prominently mentioned as holders of a large portion of the patents. While certainly not conclusive by any means, Qualcomm was noticeably absent from the list.
Editor's note: To read Part 2 of this column, in which Faulkner examines Qualcomm's issues with emerging markets and the value of its royalty stream, click here.
At time of publication, Faulkner was long Texas Instruments.
Bob Faulkner has been in the investment business for 18 years with an exclusive focus on technology stocks. Under no circumstances does the information in this column represent a recommendation to buy or sell stocks. Faulkner appreciates your feedback;
to send him an email.