(Bored engineer comments on DSPs and cell phones...)
It looks to me like the sales for DSPs are likely to be in flux, with market share moving around from company to company as relative technological advantages change. Maybe the darth vader of processor designs will come through:
Ericsson, Qualcomm call patent truce, leaving standard fragmented -- Accord clears tentative path to 3G
"This is the only market Intel has seen that will be bigger than the PC in five years," Strauss said, referring to that company's recent decision to co-design its first DSPs.
techweb.com
On the other hand, design wins for ARM said to be way up:
Claims 70 percent of this year's digital-handset market -- ARM's sitting at the core of mobile-phone boom
Nonetheless, at least one voice outside of ARM is acknowledging the company's commanding market lead in digital-handset cores. Joe D'Elia, senior industry analyst at Dataquest Europe, said he believes ARM, "surprisingly enough," has indeed reached a position of dominance. "They have got a very high level of design wins with handset manufacturers. In the last 12 to 18 months, the ARM processor has become the core that people have to design with.
...
ARM has repeatedly said that its strongest area for design wins is in the mobile-handset market, which uses the ARM7TDMI Thumb core. Mobile-phone makers hold power efficiency-to preserve battery life-at a premium, and since its creation in 1990, ARM has claimed Mips/W leadership for its processors.
techweb.com
But with the very high PEs DSP companies (like TXN) sell for, you really have, as an investor, to take a much longer look at the technology. A lot of people are working on system on chip solutions for mobile phones:
ISSCC confronts post-PC world -- Analog gurus in quest for one-chip mobile phone amid technical, economic hurdles
In a review paper, Seshadri Subbanna, who manages BiCMOS technology research at IBM Microelectronics (Hope-well Junction, N.Y.), described some of the devices his team had built using high-frequency SiGe as the npn heterojunction bipolar transistor within a BiCMOS process. The devices include ECL ring oscillators, 4-bit D/A and A/D converters with 8-GHz sampling rates, frequency dividers and active filters. These parts could be integrated with DSP and CPU cores constructed in CMOS, Subbanna believes.
techweb.com
STM pursuing CMOS for single-chip phone
Other European semiconductor makers and many universities around the world are also eyeing CMOS for radio-frequency applications, although a true single-chip phone combining power amplifier, memories and the necessary passives for today's architectures remains a distant prospect.
techweb.com
"IP" not withstanding, as systems become more and more integrated, the number of companies contributing to the end product has to decrease. Most cases of integration improvement has left some companies, or at least some products
of some companies, at death's door. The big question is who will be the survivor, the DSP makers, the handset makers, or someone else?
In all these engineering integration crunches, the important question, I think, is who is in control of the rarest and most difficult part of the design/manufacturing process. As integration proceeds, those companies are the ones that usually pick up the market. For mobile phones, my guess is that the dominating intellectual barriers to entry are possessed by the handset makers, and resides partly in their software, partly in their RF design abilities and partly in their sales/marketting/brandname advantage. For this reason, I would expect them to eventually cut the DSP makers out of the deal.
In any case, selling DSPs to the cell phone industry is risky business, and one would not want to make a bet that the established industry leaders are doomed to stay in that position.
Of course, all this is IMHO.
-- Carl
P.S. Dang! I can't believe this back-up is still going on! |
