Intel Fires Up New Atom Processors

Intel Fires Up New Atom Processors

Opportunity: 100 million chips in three years?

So, why is Silverthorne a big deal for Intel? If you look at the processor from the traditional view, it is nothing special. No new record levels of clock speeds, no multi-core and it doesn't hit a mass market - yet. But: Silverthorne expands Intel's x86 processor portfolio on the very low end to a product offering that now reaches from handhelds to supercomputers - or, as Intel likes to call it, from "Milliwatts to Petaflops". Yes, it is exaggerated, but you get the point.

Broken down to its very basics, the business model of a semiconductor company is to sell chips, lots of chips and more of them every year. It's no secret that the desktop CPU market is declining and the notebook CPU segment is the growth area right now. But growth in this segment doesn't blow off your socks in terms of unit numbers. Server CPUs is also a business that does not provide room for substantially higher CPU shipments every quarter. If Intel is looking for new growth areas, which ones are there?

Intel believes that it will be the handheld segment that will open the door to a whole new market, which has the potential to grow to a demand of dozens and even hundreds of millions of processors within a few years. Atom isn't ready to play in the smartphone market yet (which has an annual market volume of about 150 million processors by itself), but Intel wants the platform to capture market share in a segment that offers portable navigation, video and gaming. Unit sales of portable navigation devices (PNDs) were about 16 million in 2007, according to Intel; portable video players reached 10 million units and portable game consoles about 20 million. Obviously, there is a big market, which Intel thinks will expands to about 100 million units over the next two to three years.

Menlow isn't the platform that will see shipments of dozens of millions of units. The form factor and battery time of these 5" devices is in our opinion not convincing enough to attract every mom and dad to buy such a $500+ device for themselves and their children, who, according to Intel, are skyping, texting and facebooking all the time. But it is easy to see that Menlow will spark a device generation that will easily outgrow the installed base of UMPCs, which we heard stands at about 2 million worldwide.

In a way, Menlow could be powering the early-adopter version of a portable computing and communication platform the UMPC was promised to be. We have seen one promising MID, but there is no doubt that this generation will mainly bring OQO-sized devices that are too large and heavy to fit in your shirt pocket and, sorry Intel, seem to be designed for early adopters and not the mainstream consumer. That mainstream platform could be Moorestown, Menlow's successor. Moorestown is scheduled for a 2009/2010 launch and will be a SoC (that integrates the chipset) and Intel first try to become a serious player in the smartphone market.

Perhaps we are wrong and consumers will buy MIDs right away. But, on second thought, we are fairly sure that this scenario would require Apple to help design and market MID. Maybe someone should call Steve Jobs?