Those lucky sods of you out there who have purchased
our fantastic definitive market report on multimedia devices in handhelds,
who bask in the confidence and glory of now knowing everything there
is to know about the market, also know that the report predicts a consolidation
of the players as too many suppliers chase too few customers for the
same orders. Not a new phenomenon but something that everyone seems
to have to learn again and again.
The big news this week was Qualcomm’s $170 million acquisition
of Iridigm Display Corp., and the surprise of Broadcom’s $120 million
acquisition of Alphamosaic.
Qualcomm reaches out
Qualcomm’s move, in acquiring the rest of Iridigm (they
already had 14% of the company), shows they plan to own more of the
platform and see Iridigm’s super low-power MEMS interference light modulation
scheme (which they call iMoD) as the way to push handhelds to the next
level—i.e., larger, higher resolution, screens but without the
battery load that are less expensive to build than LCDs. With the inclusion
of color displays in all types of phones, including models at the low
end of the market, the cost of the display has become an even more significant
driver in the overall cost of the handset. An iMoD display should cost
significantly less to manufacture. Qualcomm is now in position with
the recent ATI deal to offer end-to-end solutions for hand-helds. This
is very cool stuff, but equally important, it shows how the big boys
will dominate the market.
And Broadcom takes some marbles home, too
Broadcom is also big, and no newbie to the handheld market.
The company made its first move in April when it acquired Sand Video,
a video compression company, for $77 million. Broadcom withdrew Sand
from the market and assimilated it. Broadcom currently provides cel-lular
OEMs with EDGE/GPRS/GSM baseband processors and is now sampling WCDMA
baseband processors as well. The company was first to demonstrate high-speed
“four-slot” EDGE performance at greater than 200 Kbps, and
they recently acquired Zyray Wireless (for $96 million), a developer
of WCDMA baseband co-processors, which enable up to 384 Kbps of wireless
Now Broadcom has picked up one of the more innovative
companies in the media processor biz, Alphamosaic (see last week’s issue
of TechWatch, Sept. 6, for a description of their VC02 media
processor). Alphamosaic’s products are offered as discrete co-processors
and will also be integrated into Broadcom’s cellular baseband and application
processor roadmaps in the future. Alpha-mosaic’s first-generation VC01
multimedia co-processor is currently being used by Samsung, and we expect
them to pick up the VC02 and some other products coming out of Alphamosaic’s
lab, which we saw a couple of weeks ago. Alphamosaic is doing about
$8 million in sales this year, and had expected that to skyrocket to
$30 to $40 million next year.
More consolidation will happen, and we even have a few
clever guesses about who—I’m sure you do, too.
A new market segment—HVP
ATI has done a clever thing and created a new market segment,
for which we’ll have to come up with a name. They’ve built a chip that
uses the bandwidth of PCIe to allow use of the system memory for graphics
memory, much like an IGP does. It’s a tweeny—between an IGP and
a stand-alone or dedicated VPU, and they call it HyperMemory. Hybrid
might be better, and maybe that’s what we’ll call it, an HVP.
Intel’s new Extreme Graphics 3 is a kick-ass part that
can run DirectX 9 games, and is a challenge to market leaders ATI and
Nvidia. This HVP offers a lot more power for a fraction more in cost
and could blunt some of the gain Intel will make with the new Grantsdale-G.
If the concept gets any traction I expect to see Nvidia,
S3, and XGI come out with similar offerings.