Samsung, Nokia, and Qualcomm and CDMA Chipsets and Handsets
The mythology of the Qualcomm threads dictates that Samsung will only use their own chips on the low end. The 3rd most profitable IT company in the world can sneeze at Qualcomm in that regard and so can the 4th largest. Stabilize the initial chip and work up the ladder.
Slacker,
<< I think you oversimplifying the move from the low-end of the CDMA handset arena to the high-end. >>
Perhaps.
I think many of my colleagues overexaggerate it.
<< In the end, I think Samsung and Nokia will need to tier their chipset offerings in a similar way as Qualcomm. This is definitely possible for both companies....but this is going to take time AND money. >>
First, in order to examine this, I think we have to differentiate between Qualcomm, Nokia, and Samsung. We really have 3 different very successful fruits to compare. Two very big fruits and a smaller one. They all have time, AND they all have money.
Qualcomm is a fabless OEM wireless IC manufacturer with exceptional design capability, unquestioned competence in CDMA ASIC design, and proprietary control of open architecture in CDMA2000, which gives them a real first mover advantage, although perhaps less now than in the past. They have adequate cash to fund very good R&D.
Nokia by contrast is a handset manufacturer not an IC manufacturer. They manufacture whole product rather than a component. They do however, have a world class strategic partner for IC manufacturing, several smaller highly competent partners, and they have considerable proven core competence in IC design and obvious economies of scale. While apparently advancing nicely, they have not yet totally mastered CDMA2000 design or the idiosyncrasies of the CDMA market (many more carrier specific one offs than in GSM/TDMA). They'll be the first to admit that and they do. They are of course handicapped by the fact that Qualcomm is the proprietor of the CDMA2000 open standard. They are further handicapped by the fact that they don't manufacture CDMA infra and were handicapped by the fact that they were not hands on involved in the initial implementation and debugging side by side with Qualcomm of 1xRTT in Korea which gave Samsung and LGE a major head start. Offsetting this somewhat is the fact that they don't have to worry about how their own chipset design endeavors are viewed by Qualcomm since they have no strategic partnership with Qualcomm and have always designed their own chipsets.
Samsung by contrast again is an electronics super power that is not as focused as Nokia or Qualcomm but has some major advantages in diversity to offset that. Despite the fact that they are fabbed and manufacture IC's they are not (yet) as advanced as Nokia in wireless IC design, but I think that's changing rapidly and they do have excellent IC partners outside of house as well as in. They are handicapped by the fact that at the same time that they are trying to master CDMA2000 chipset design they also have to master WCDMA and EDGE and are well behind Nokia in this regard, and are more reliant on Qualcomm as regards WCDMA and other partners for EDGE. They also have to be somewhat careful of their relationship with Qualcomm while they are in the crawling stage before they can walk like Nokia already is, or sprint like they probably eventually hope to do.
<< There is a reason for Qualcomm's tiered chipset strategy.....it provides far greater efficiencies than a one-size fits all approach. >>
That's exactly like what Nokia has mastered in GSM handsets with their segmented approach which marries with their platform approach (both concepts they pioneered) and appear to be doing now with CDMA handsets as they work up the ladder starting with a market sweet spot.
Qualcomm needs a tiered chipset strategy badly. To their credit and starting with the MSM5105/5100 they have one (where they used to offer the previous version or two as their low end). The major issue this will help to overcome is the cost differential between a GSM handset and a CDMA handset now that CDMA is increasingly destined for emerging markets.
One thing I wonder about is how the broader range of segmented offerings will play out with manufacturers who will design software around this chipset or that chipset, but that is a whole separate discussion.
<< We don't know the spec's for either the Nokia or the Samsung chipset, so it is tough to look at exactly what problems they might encounter. >>
We will never know specs on the Nokia or Samsung chipsets, unless they decide to become a chipset OEM. What becomes important to now is what functionality the finished product has, at what ASP, and at what (overall) margin. Nokia pulls back the kimono on this just slightly and on rare occasions like strategy updates, but they play this very close to the vest. Samsung essentially does the same thing on their wireless side as opposed to the IC side which is a different ball game.
<< The one problem that has been repeatedly mentioned....a low-end chipset that isn't a zero IF solution is going to lose a substantial part of its price gap with Qualcomm's chipsets. >>
If you have to pay Qualcomm's highest in industry margins, yes. If you don't, not necessarily. Meanwhile Qualcomm claims to be the only industry player to apply ZIF or N-ZIF to CDMA chipsets as opposed to GSM. Are they? Probably. For how long? Who knows. Nokia (or Samsung) aren't about to reveal their chipset architecture (although their are houses that will dig into this and break out ... for a hefty price).
Qualcomm has romanced the heck out of ZIF for (all to long) a time. As they should. It plays well for us shareholders and hopefully will help them maintain a competitive edge, good market share, and good margins.
<< A low-end CDMA chipset is unlikely to have the substantial number of integrated features that Qualcomm's mid to high-end chipsets do. You can offer every single one of those features using additional chipsets around an initial low-end chipset....but at substantial penalty in size, battery life and cost. >>
The proof of this pudding will be what size, actual battery life, and cost, is available relative to price and margin from Nokia, or Samsung at a given point in time. It's a little early to tell with Nokia and even earlier to tell with Samsung and this applies to both CDMA2000 and WCDMA.
Indications are that the Nokia 6650 has a reasonable degree of integration and a reasonable feature set although not as fully featured as some of its competitors despite being perhaps more tightly integrated. We won't really know until we see this product shipping commercially in volume. I expect considerably more out of their next generation chipset that will power their next more powerful model releases.
Same with their CDMA2000 product line, where they are still working on incorporation of carrier specific requirements in software. They obviously are gaining real confidence in the base platform as it currently exists and consider it to be highly scalable. I suspect it is, although it is really to early to tell. I don't consider them out of the woods yet.
<< Outside of the above features is going to be the need (perhaps?) to provide 1xEV-DO in their upcoming chipsets. This will require another round of qualification with vendors and operators. >>
This is probably (possibly) more applicable to Samsung than Nokia since Nokia won't be marketing CDMA2000 in Korea or Japan. Samsung might attempt it, but they may rely totally on Qualcomm. 1xEV-DO is more proprietary and less open than 1xRTT Release A, which is less proprietary and less open than 1xEV-DV. I personally don't think Nokia will ever incorporate DO, nor do I think they need to meet market share targets.
Samsung is harder to evaluate than Nokia for any number of reasons. Qualcomm may be their primary chipset supplier for a long, long time. I sure hope so because I hold Qualcomm and don't hold Samsung. There is some risk attendant to how Samsung's possible migration to their own CDMA chipset occurs, but it is not a migration that will occur over night, by any stretch of anyone's imagination and I think the sharper analysts have this pegged pretty good.
Those are my thoughts. Obviously there is liberal opinion stated here, but it is based on considerable time spent carefully observing Nokia, less so with Samsung, and plenty of time observing both Nokia and Samsung in 3GPP and 3GPP2. Still it's just reasoned opinion.
All JMHO and FWIW.
Best,
- Eric - |
