Who's eating whom?

A lot is being made in the tech press of this deal between Nokia and Intel, and how it surely represents Intel eating into ARMs market share.

Let's put things in perspective here.

In Taipei (Computex 2009), we saw countless ARM based chipsets being touted by companies like Qualcomm, NVIDIA, Freescale, Samsung, Texas Instruments and others, each boasting a sizeable number of Original Device Manufacturers (ODMs) . These weren't just any slouches, either - Asus, Acer, Pegatron, Inventec etc. Household names in many cases, and typical PC industry stalwarts. These are companies who have been manufacturing laptops forever, or are breaking into the game in a big way. And here they all are, one year on, allowing chipset vendors to display devices bearing their name. Sitting in those booths were missed-design wins for Intel; capacity in factories diverted into manufacturing ARM-based equipment, displacing Intel.

If you like, Taipei was the show that really splattered it on the wall for Intel. They finally saw that their dominance of the computing market is at last being challenged, and all at a time when the world is turning sour for them with one of the largest anti-trust judgements in history levied against them.

I found the following blog on ARMs website, which has some pretty interesting dissection of how the current computer manufacturing business could be changed if choice were introduced into the market place for these ODMs in taiwan.

But away from the ODMs and onto Intel again, who continually claim that standardisation is what they'll bring to the mobile device market. They claim that if they standardise the platform on which mobile devices are made - and by the way, be the only supplier of that platform - then OEMs won't have to work as hard at differentiating their devices. This is tantamount to the death of innovation in the mobile space. If the arguement is purely based on the fact that all of the peripherals in the SOC that they lay down will always be in the same part of the memory map so that an OS will always know where to find them, this arguement doesn't hold any water.

All OSes have layers of software that abstract the underlying hardware (a HAL, or hardware abstraction layer) so that they don't need to care where the graphics processor is, or how much memory it has etc.. Intel are dreaming with this arguement, since the countless ARM silicon providers all have HALs for their given SOC platforms on a number of OSes - it isn't something an ODM or OEM needs to care about. So on that front, Intels arguement may seem clearcut to Intel - it merely requires that every other processor architecture and platform disappear, so that there is only theirs left standing!

With the Atom cutting into the sales of their higher cost laptop processors - adding confusion to the market - and with ARM devices eating their market share come the end of this year, I predict a great fall in Intels profitability at the end of 2010. ARMs business model is less volatile than Intels, and the analysis suggests that - for now - it is Intel who is getting eaten, both by itself, and by ARM.