San Francisco?s Moscone Center was home to Intel?s annual Developer forum this past week. It was fairly low key with a couple of 3.1 Richter Scale announcements. A little peaking under the surface reveals that Intel is repositioning their consumer side to address changing market demands?
Intel is a margin driven organization with sales that roughly divide into two categories consisting of corporate [33%] and consumer [66%]. Corporate sales have, for the most part, maintained good volume and margin though the consumer side of the business seems to have suffered from a shift in consumer spending habits. Consumer expenditures for desktops and high end notebooks have shifted to netbooks, nettops, tablets and smart phones, a market segment that?s quickly approaching commodity status along with lower margin expectations. Then again, Intel is all about the volume and in mobile phone business, there is a potential to ship more processors than Intel ships today. Lower margin yes, but if the sales is measured in hundreds of millions, the bottom line would increase from a few hundred million to a few billion dollars extra.
Intel began investing in the mobile segment beginning with the Atom processor development in 2004 in an effort to address the needs of what the company calls the "Mobile Internet Device" market. A not so obvious, but apparent strategy, was the company?s desire to retain and strengthen their hold on the pervasive "Intel Architecture" model within the mobile segment. Intel?s involvement with the ARM processor effectively ended when it sold its XScale acquisition to Marvell [~US$600 million] in 2006 leaving the Atom processor line as their sole roadmap in the segment. Marvell went on to develop an ARM-based processor series now widely used in smartphones.
Although the Atom processor is relatively low power for an x86 architecture processor it requires chipset support driving up power levels above those of competing ARM implementations. Despite this, Intel has captured a wide number of netbook and nettop design wins with the codenamed "Pineview" Atom processor which combines memory and graphics control on chip but requires a Platform Controller Hub [PCH] codenamed "Tiger Point".
Introducing Atom E600 SoC
Intel has been working on the Atom series processor for six years and has yet to field a System-on-Chip [SoC] version fueling speculation of what intention?s the company might have as regards the embedded control marketplace. The E600 announcement permanently put to bed any further speculation of the company?s intentions. This is a larger undertaking than many realize and is one of the more interesting system level development efforts that Intel has ever entertained. The company?s strategy is to bring system level integration costs down while spreading the Intel Architecture into new application areas including in-vehicle infotainment systems for cars, smart grid devices and IP media phones.
Intel Atom E600 "Stellarton" – Intel packs Altera FPGA with an next-gen Atom core
One processor detailed during a keynote is codenamed "Stellarton." It consists of an Atom E600 processor coupled with an Altera FPGA on a multi-chip package that provides additional flexibility for customers who want to incorporate proprietary I/O or acceleration. Although Stellarton provides customers with a fast time-to-market the question remains whether Intel will entertain a full-custom path for high run devices? We suspect this is a trial run for the future integration of processing units from the TeraScale and "many-core architecture" projects – formerly known as Larrabee. We might be wrong, but that integration would be the traditional path that Intel follows.
The E600 represents a fairly significant departure from Intel?s normal business regimen. There?s concern that the company is not well prepared to handle support requirements that a significant number of design wins would have on the company, on the other hand, the company could just as easily spin it off as a separate operating entity. One thing is clear, Intel needs this capability to capture, service and retain top to bottom processor slots at their large OEM customer accounts.
Sandy Bridge is the result of a historically rocky road in Intel?s efforts to achieve market acceptable graphics performance on Intel Architecture silicon. Sandy Bridge is Intel?s mainstream architecture for 2011. It won?t enter the value space till sometime in 2012 with first units shipping in early 2011 [April?].
The graphics engine design was physically "lifted" from supporting chipsets and placed on the CPU for the first time. Known as Intel integrated graphics it has been shipping in a number of notebooks for approximately six years and also shipped in a significant number of PCs sold worldwide. Design win announcements are expected to be announced at the Consumer Electronics Show in January.
Opher Kahn, Senior Principal Engineer of the Sandy Bridge design team, said "We?re not trying to target the most high-end discrete [standalone] card. We don?t have the bandwidth, we don?t have the power budget. We?re trying to do the best experience for the mobile platform."
Sandy Bridge fits within Intel?s plan not only the mobile space but that it lowers the overall power budget regardless of end use. Intel has engineered a very competitive solution for today?s market requirements and conditions.
Wireless & Security
Intel?s recent acquisitions of Infineon?s wireless group, Texas Instruments cable modem group and McAfee are flashing red indicators of the company?s future intentions in the smartphone market, heightened overall internet connectivity and internet security respectively. This will pit the company against the likes of Qualcomm Inc. and Broadcom Corp. Intel is now executing their vision of the company?s future. It looks fairly Mobile?