Subscribe to our Feed

Texas Instruments: Wireless Woes

Posted on Thursday, Mar 27th 2008

By Vijay Nagarajan, Guest Author

In the last two parts of this series, I looked at TI’s position in the analog and DSP markets. We also concluded that the OMAP was central to TI’s wireless strategy. I am afraid that TI has put all its eggs in this one basket, and has perhaps miscalculated the strengths of its competition, especially the new entrants. And the result? It has all but lost its dominance in the wireless market and also the golden chance to make the best of the mobile revolution.

Before we look at the outlook for TI’s wireless business, let us step back to look at some of its current statistics. About $2bn of TI’s 2007 revenues were from 3G with the 3G OMAP alone accounting for $500mn. Custom chips that TI manufactures for the likes of Nokia accounts for most of the rest. Let me reiterate again now that these custom-chips, that make TI the largest 3G chipset vendor at more than 50%, do not have its IP for the baseband. For more relevant statistics, I will direct the interested reader to the earlier articles of this series here and here.

Also, TI has consciously stayed away from developing 3G digital baseband. I think that this is a mistake. I believe that baseband is the most important part of the 3G chip. This contrasts TI’s position that the application processor is the heart of the phone. I do agree that 3G is about the experience – the music you hear, the TV you watch, the games you play, the high-resolution still images and videos you capture are all integral to tomorrow’s phones and will be enabled by the application processor. But what good is it if your mobile data and internet experience is bad? Unlike voice, data will hog bandwidth and limit the network’s capacity effectively reducing user throughputs. So it serves the mobile user well to get a phone with a good baseband processor with well-implemented advanced receivers.

This segment is intended to add context to the next two parts where I discuss TI’s wireless strategy. It is important that you digest these points before proceeding so you can get a holistic perspective of TI’s position in wireless.

This segment is a part in the series : Texas Instruments

Hacker News
() Comments

Featured Videos


[...] pushing merchant chipsets for current third-generation wireless protocols, something fellow blogger Vijay Nagarajan has pointed out in a series of thoughtful posts. Essentially, TI is getting out of an increasingly competitive [...]

TI to Sell Part of Its Wireless Chip Biz - GigaOM Monday, October 20, 2008 at 3:37 PM PT

[...] In our deep-dive analysis of the company, guest author Vijay Nagarajan pointed out that TI had a flawed baseband strategy in which it wasn’t investing to develop a 3G digital baseband. [...]

Cheap Chips: TI, Broadcom, STM - Sramana Mitra on Strategy Monday, November 3, 2008 at 9:58 AM PT

TI winding down its merchant baseband business has been widely discussed. A big blow seems to be Nokia’s multi-sourcing strategy. TI winding down this division at a time when the trend has been to move away from R&D intensive ASIC parts to more prevelant ASSPs is something to mull over. As is the case with such strategic changes, the effect of this decision will be seen over time. In fact, although Nokia announced it’s multi sourcing strategy in 2007, the proportion of TI’s revenue from Nokia increased from 19% in 2007 to 20% in 2008. It should be noted that total revenue fell in 2008 over all.

TI management takes pain at every possible opportunity to stress that the same multi-sourcing strategy that would hit TI in areas where it was/is a dominant supplier would benefit TI in other areas where it is not a dominant supplier. analog and power management parts? There is some truth to it, although the deal making power without the accompanying merchant chips will be under a lot more stress.

Other players in the market will be working hard to make the best of this transition time at TI. Toshiba already edged past TI to claim the 3rd position among top semiconductor suppliers by revenue for 2008. STM is seen poised to take the second position -currently held by TI- in top cellular baseband processor suppliers for 2009.

Will Strauss of Forward Concepts pointed the trend in providing application processors combined with baseband processors. If this catches on, TI might loose some design wins for lack of a baseband solution to package with OMAP processors.

Google’s Vic Gundotra said, in Mobilebeat 2009, web based apps is the future. There are many advantages to this standardized platform approach. This has to be taken with a pinch of salt though as Google stands to gain a lot if users spend more time on a browser than on a native application. However, if this trend takes off, application processors will play an increasingly important role. TI’s OMAP platform with top of the line ARM cores and integrated connectivity solutions will stand to benefit.

TI has a complete suite of integrated connectivity solutions and a strong presence in application processor market. So long as device designers do not see a pressing need to further integrate baseband processor, connectivity solutions and the application processor, TI can play its cards to its advantage. Samsung and Marvell – top 2 and 3 among application processor suppliers are not sitting idle and are working hard to grab some market share from TI. Samsung announced developing 4G chips. Marvell has Tavor.

Nalini Kumar Muppala Monday, July 27, 2009 at 3:57 PM PT

Good analysis, Nalini. Do you have a sense of how the design wins in the application processor market break down between TI, Samsung and Marvell? What are the precise market shares and design wins?

Studying that would give us a reasonable idea about the direction of this business for both TI and others.

Sramana Mitra Monday, July 27, 2009 at 4:15 PM PT

Btw, one other thought … the smart phone market is driving a lot of the growth / margin in the chip business at the moment. But for the growth to continue beyond the early adopter / early majority segments, the phone prices would need to drop. And in any case, we’ve already started seeing the price wars in handsets. For the emerging markets to access smartphones would definitely require prices to drop significantly.

What impact does that have on an integrated baseband versus a fragmented baseband strategy?

Sramana Mitra Monday, July 27, 2009 at 4:20 PM PT

According to Forward Concepts, the Stand-Alone Application processor market totaled $1.3B in 2008. TI dominated with 43% of the revenue, followed by Samsung with 17.4%, Marvell with 17.1%. Marvell shipped more units than Samsung but at a lower ASP. ST-Ericsson took 9.4%.

Major phone design wins :
TI : Palm Pre, Nokia E, N series, and many more.
Samsung: Apple iPhone, Samsung Jet, (more?)
Marvel: several Blackberry models, Palm Treo, Motorola Q.
2007 design wins include such familiar names as Asus, Blackberry, Fujitsu/Siemens, HP, Hyundai, Lenovo, Motorola, O2, Palm, Samsung, Sharp, Vodaphone

Interesting points:
Marvell sold to Samsung as recently as 2008-02-05. As Samsung development matures, Marvell will not see such dwins.
Marvell seems to be gaining from smartphone growth in China.
Several Nokia phones use TI OMAP 1510, 1710, 2420 upto N95. but the latest N series additions N96, N97 use ST Nomadik. Nokia seems to extend the multi sourcing strategy to application processor parts as well.
Nokia announced an alliance with STM at MWC 2009 to provide devices based on U8500 platform: ST-Ericsson’s application processor + HSPA Release 7 modem.
If ULC+Application processor model were to take off, Samsung will gain from its alliance with Infineon.
Infineon and Samsung started collaborating on smartphone reference design as early as Feb 2003. This relation ship lead to fruition in iPhone.
Samsung along with Intrinsity is pushing Cortex-A8 performance to 1GHz.
Growth beyond phones:
TI revamped OMAP product line, which was originally targeting handset market, to address other markets as well: PND, embedded gaming, portable medical equipment, etc. A lot of push into MIDs.
OMAP3503 in M2M Solution’s Homebox for surveillance.
Samsung targets Mobile computing (PND, PMP, etc) separate from Mobile communication.
Samsung in Chumby digital photo frames, Navico’s GPS devices.
More than just smartphones, Marvell seems to be targeting other consumer devices such as connected digital photo frames, portable navigation devices, portable media players, automotive dashboard displays and infotainment systems, home automation, portable TVs,
Integrated Modem+Application processor approach:
With it’s integrated approach,QCOM stands to gain if ULC’s start adding more smartphone functionality.
Infineon+STM+Nokia collaboration, Infineon+Samsung application processor for Nokia phones is a possibility too.
TI will have to workout an alliance.

Nalini Kumar Muppala Thursday, July 30, 2009 at 4:19 AM PT