2 Comments

Summary:

[qi:gigaom_icon_chip] Texas Instruments last year said it would exit the wireless baseband business (it will still make custom radios for clients, but will dump its catalog of wireless baseband chips), and today the Wall Street Journal notes the effect this is likely to have on TI’s […]

[qi:gigaom_icon_chip] Texas Instruments last year said it would exit the wireless baseband business (it will still make custom radios for clients, but will dump its catalog of wireless baseband chips), and today the Wall Street Journal notes the effect this is likely to have on TI’s earnings. The consensus is that with TI dumping radios and betting on applications processors and analog semiconductors, baseband revenue will fall faster than its new core lines of business can grow — leading to a potential gap in revenue and profits. Wireless baseband chips currently make up about a fifth of TI’s 2008 revenue.

These chips are rapidly becoming a commodity, and so it makes sense to cede the market to larger players like Qualcomm and MediaTek, but TI is still placing a huge bet that it can make up the lost revenue on sales of its applications processors, analog chips and digital signal processors. I summed up discussions I had with the company in March when Greg Delagi, head of TI’s wireless terminals business, argued that TI thinks smarter phones, with its ARM-based OMAP applications processors designed for mobile Internet devices and netbooks, will win. But it’s possible that instead of the high-end application processors that TI is building, vendors may elect to build consumer devices with highly integrated chips that the company won’t be able to churn out profitably.

As for analog, most such chipmakers like National Semiconductor, Analog Devices  and others are expected to do well since these chips are essential when it comes to converting the real world to the zeros and ones required by the digital world. Analog chips are like translators, and now that everything from my tire pressure to my TV channels are regulated by digital processors, those translators are needed everywhere.

Texas Instruments has made big leaps of faith in technology before, getting into the digital signal processing business in the late ’90s as the rise of broadband connectivity everywhere meant an increase in the number of packets a router needed to sort. DSPs are the chips that handle packet routing and other jobs that need a high level of mathmatical processing in a hurry. They are also in home networks and cell phones handling the influx of digital information.

You’re subscribed! If you like, you can update your settings

  1. Gadget Sleuth Monday, July 13, 2009

    “but TI is still placing a huge bet that it can make up the lost revenue on sales of its applications processors, analog chips and digital signal processors”

    A bet it will lose, if Intel is having trouble doing this sort of thing. I actually TI had went under awhile back; it seems like i’ll turn out to be right.

  2. “DSPs are the chips that handle packet routing and other jobs that need a high level of mathematical processing in a hurry.”

    Packet routing is actually a good example of a function that is better doen with dedicated hardware and does not have a ‘high level of mathematical processing’. This is in marked contrast to a cellphone or home metworks where key baseband functions are handled by DSPs including OMAP chips.

    As for Gadget Sleuth’s comments above, I don’t think TI will go under at all; their revenue will go down but profitability should improve (when I was briefly at TI in the mid-90s, the HPA division (High-performance Analog) produced most of the profit for the company). And Tl lead the industry in analog sales; more than National, ADI, Linear Technology or Maxim.

    So my bet is the other way. Intel is not the best example as they have never succeeded in communications, particularly RF. Their radios and mixed-signal expertise tends to come from acquisitions; in any event the have struggled in the non-core-processor area — one examplei is 802.11a where they tried but could not makes the WiFi radio and ultimately used Atheros in the first generation of modems.

Comments have been disabled for this post