Qualcomm, Inc.
Re: 50% of WCDMA?

Format for Printing

Format for printing

Request Reprints


By InCards
March 8, 2004

Posts selected for this feature rarely stand alone. They are usually a part of an ongoing thread, and are out of context when presented here. The material should be read in that light. How are these posts selected? Click here to find out and nominate a post yourself!


<< Would love to hear (make that read) what you folks actually think about this projection of QCOM acquiring 50% of the WCDMA chipset market. >>

I'll repeat what I have stated on this board and others several times before in pretty much the way I have always stated it.

I'll be delighted if Qualcomm can capture 25% of the nascent WCDMA chipset market, and would consider anything beyond that a bonus.

I also think that their 50% target, set almost 2 years ago without a timeframe attached and reiterated several times by Dr. Irwin Mark Jacobs, Tony Thornley and others, is overly aggressive, and even unrealistic.

That said, I'd be very disappointed if Qualcomm set a wimpy goal of 25% or 30%. If a company doesn't set a stretch goal, what they achieve is going to wind up being less significant than what they hope to achieve, and I think that each and every member of the Qualcomm team, top to bottom, believes that they can achieve their stretch goal even though they realize it's a stretch, and that's the attitude I want in management and employees of a company I invest in.

I firmly believe that Qualcomm is going to be a major player in 3GSM WCDMA chipsets. They engaged late in active 3GPP participation although by virtue of their ETSI membership, which dates from November 1997 they have been a 3GPP participant from its inception. They also started 3GSM WCDMA chipset development considerably later than half a dozen others, but once they fully engaged they did so with vigor. One of the things they did that impressed me was to develop their own GSM GPRS protocol stack in house rather than contracting and licensing outside, which they originally planned. I believe that mastery of the complete end-to-end technology is very important.

<< If TI and associates produce and, or market a WCDMA chip by the end of 2004 that may put them about 9 months behind the curve. >>

It is not a matter of " If ".

Nokia has two commercial 3GSM WCDMA handsets currently delivering using their own custom designed chipset featuring a baseband designed in collaboration with TI. The first (6650) was type approved in Europe and Japan in October 2002 and has been delivering since, and in low commercial volume since mid-year last year. The 2nd model (7600) which features considerably higher integration, more features, reduced size and weight, and increased battery life has been delivering since January. Expect more announcements shortly and expect more features and integration.

I am relatively sure (but not positive) that Ericsson uses TI for the baseband of their U100 chipset platform that Sony Ericsson will use in their Z1010, which should commercially deliver shortly and Sharp will evidently use the U100 in their Sharp V801SH which will deliver for Vodafone Live! on Vodafone KK, first and then elsewhere.

TI is currently sampling their TCS4105 direct conversion four-chip dual-mode chipset for both the WCDMA and GSM/GPRS communications protocols which supports Class 12 GSM/GPRS and 384 kbps WCDMA services, and development on this platform has been taking place by several OEMS and ODMs for ~2� years using prototypes of the platform and TI design tools.

NEC, Matsushita (Panasonic), and Fujitsu, the major suppliers to DoCoMo FOMA, all use their own baseband silicon coupled with OMAP processors by TI. The newest FOMA 900i Series which are rather sensational handsets and which are garnering oohs and ahs from all corners of the globe, use the latest TI OMAP processors. They are the pacesetters but they are not multi-mode.

Two weeks ago in Cannes TI released their ARM11 powered OMAP2 multi-engine parallel processing (ARM1136 at 330 MHz, TMS320C55x� DSP at 220 MHz, and 2D/3D hardware MBX/VGP accelerators, up to 2 million polygons/sec) family of applications processor family. It's pretty powerful stuff. TI expects these products to be in shipping 3G handsets by year-end or more realistically Q1 2005. Here is the release brochure for OMAP 2:

I recently clipped 9 slides from TI's OMAP2 launch Webcast to 3 posts beginning here:

I look (perhaps overly optimistically) for Nokia to have the first commercial 3GSM WCDMA handset powered by an ARM11 processor, just as they were first with ARM9, and (I think ARM7), but probably in a custom design, rather than using OMAP 2, and the Japanese may scoop them on this go round.

We should be getting visibility shortly from Qualcomm on when they plan to sample the ARM11 powered multi-engine 7000 series, but I'm guessing commercial product shipping in handsets is at least a year out, maybe 1� years. In the interim the MSM6250 now sampling looks like a very fine product and will power the initial releases by Samsung, LGE, and others, and the MSM6275 which is yet to sample will incorporate the HSDPA extension of WCDMA.

A prime differentiator of TI v. Qualcomm other than the fact that TI owns their own fabs, does their own packaging, and is very advanced in process design, is that they do custom designs for companies with in house development, and they also sell a variety of independent module components that interact with others chipsets, and along with ARM, ST Micro, and Nokia are driving open applications interface architecture in the MIPI alliance who now has over 40 members, with Motorola and Intel on the MIPI board. They are also the primary platform (along with Intel) employed by licensees of Microsoft's MS Smartphone OS, and provide chipsets for the majority of Symbian OS 2.5G devices shipping, and it's important to note that one Symbian 3G product is shipping (Fujitsu) and Motorola recently announced a Symbian based UMTS smartphone.

I put some emphasis on TI here in this post not only because you mentioned them specifically, but also because I firmly believe that TI will be Qualcomm's primary short, medium, and long haul competitor either with custom designs (Nokia, Ericsson, etc), highly integrated off the shelf application processors coupled with others baseband (NEC, Matsushita, Fujitsu, etc.), or fully integrated chipsets for OEMs and ODMs. I feel it is always wise NOT to underestimate competent competitors.

Those interested in exploring the TI competition in more detail might want to review TI's 56 page Wireless solutions guide:

TI's wireless chipset UMTS solutions are described here:

[Disclosure: I am NOT a TXN shareholder and never have been, but I do hold ARMHY and NOK in addition to QCOM]

Qualcomm without question has exceptional chipset design competence, but so do several of their competitors. This is a pretty crowded field they are entering, and they lack the architectural control they enjoy in CDMA and as a consequence have to react to tremendous network dependencies, to quarterly updates to 3 different release levels of the 3GSM WCDMA standard and 3 different quarterly release levels of the GSM EDGE standard rather than proact as they do with CDMA2000, where they have frequently commenced development of CDMA chipsets as much as 3 years before TIA or 3GPP2 standards are finalized. That competitive advantage is not in play when open architecture is developed in committee.

This is not a 'Gorilla Game,' it is royalty game (using Geoffrey Moore's terminology), because in the game he plays, a gorilla must have architectural control of a proprietary open technology not just IP in that technology, and a committee based open architecture is the antithesis of proprietary open architecture.

There probably will not even be any Kings in the 3G chipset game (the market leader with ideally at least 2x the market share of the next nearest competitor according to Geoff), only a few Princes, and I do expect Qualcomm to be a standout Prince.

NEC and Motorola were first to market with multi-mode 3GSM WCDMA handsets using chipsets of their own design. I expect each to have meaningful market share with chipsets they use internally or chipsets they OEM to others. At best they will be Princes, and I wouldn't even count on that.

Amongst the challenges Qualcomm will be facing is the incorporation of other RF technologies into their chipsets including in the near term EDGE, Bluetooth, and 802.11 WiFi, and in the medium term 802.16. Some of their competitors are considerably ahead of them in this regard.

The following is an alphabetical list of the major Qualcomm competitors I have currently identified who already are, or who will provide baseband ICs, application processors, or in some cases turnkey chipset platforms and reference designs for J-UMTS or 3GSM WCDMA UMTS:

- EoNex
- Ericsson Mobile Platforms (EMP)
- Fujitsu �
- Hitachi �
- Infineon
- Intel
- Matsushita �
- Mitsubishi �
- Motorola �
- NEC �
- Samsung
- STMicroelectronics
- TI �
- Via Technologies

Commercial handsets containing their chipsets currently shipping

It should be noted that several competent companies like TTPCom have developed GPRS EDGE WCDMA protocol stacks and are licensing their stacks to competent IC manufacturers like Intel, making it easier for them to enter the game. The above list will probably grow before it shrinks.

Don't count late entry chipzilla Intel out. It is not wise to overlook the fact that 3G pioneer DoCoMo and Intel are co-developing chipsets for 3G handsets.

<< I realize of course that such a discussion would be highly speculative. >>

Yes, to be sure it will be, and I personally think it is going to be very difficult to get a fix on what Qualcomm's market share will eventually be when the 3GSM WCDMA game is in full swing out about 2007 and going forwards, for some time.

Early winners may not be late winners. Vodafone CEO, Arun Sarin was recently quoted by Mobile Communications International (February 2001 issue) as saying this about their potential 3GSM WCDMA UMTS suppliers:

"Who we start with, who we end with, and who we are with in the middle will all change."

By the end of 2005 headed into 2006 some of the dust will have settled and some meaningful estimates can be made, IMO.

On January 14, Brian Modoff (BM), the analyst covering Qualcomm for Deutsche Bank Securities Inc., North America, gave his estimate of Qualcomm's potential market share in the short to medium term. I thought his estimate was too low for this year but I just realized that he was then, and is now, stating FY and I assume he means Qualcomm's fiscal year. He updated that forecast positively yesterday, based on updated intelligence.

Below is Brian's original forecast, his updated forecast, and my initial WAG which (CY in my case not QCOM FY like Brian's) remains unchanged despite what Brian estimates:

        BM Handset    BM QCOM Chipset     BM QCOM Chipset   My QCOM Chipset
         Forcast �    Share (1/14/04)     Share (3/05/04)   CY Share WAG
FY2004     16.457m        3.8%               14.1%             19.0%
FY2005     33.456m       13.1%               19.6%             22.0%
FY2006     56.668m       25.0%               30.0%             23.0%
FY2007    107.257m         -                   -               24.0%
FY2008    167.131m         -                   -               25.0%
� BM for Deutsche Bank evidently updated this forecast yesterday

It should be noted that my respected colleague on multiple wireless message boards, Jim Mullens, reported elsewhere that on a videocast yesterday that I didn't view, that Brian Modoff stated [the bold for emphasis is Jim's not mine]:

"The companies (QCOM) market share in WCDMA chipsets may actually approach their stated goal of 50% versus our current forecast of 30% in 2006."

ALL the liberal opinions I have stated above are JMHO and they are offered simply FWIW. The facts as stated are as accurately as I can make them.

I'm very interested in hearing what others here have to say in response to your timely question and their rationale for saying it. Let the discussion begin. Whoops! I see it has begun.


- Eric L. aka InCards on TMF -

Become a Complete Fool
Join the best community on the web! Becoming a full member of the Fool Community is easy, takes just a minute, and is very inexpensive.