Analyst: ARM to surpass Atom, in 55% of netbooks by 2012

Intel’s netbook dominance has yet to be challenged in a serious way, but at least one business analyst thinks that’s going to change in the next 2-3 years. The netbook market is extremely new; Atom itself isn’t even a year old—if the IT industry were to break with the existing Intel+Microsoft model, the emergence of a new product type combined with a deep economic recession could be the perfect opportunity to do it.

Analyst Robert Castellano from The Information Network believes future netbooks based on the upcoming ARM Cortex-A9 architecture and running Linux could create a market for netbooks at price points Intel and Microsoft simply won’t be able to match. Not only would the multicore A9 be cheaper than a 2012 Atom (at least, according to Castellano), the assumed ubiquitousness of cloud computing would supposedly eliminate (or almost eliminate) the need for local storage.

Not everyone agrees with Dr. Castellano’s prediction. Last week, EETimes ran a story on IDC’s own forecasts for the next 2-3 years, which differ significantly from what The Information Network believes is likely. Analyst and vice president of global semiconductor research for IDC, Mario Morales, doesn’t see x86-alternatives capturing more than 10-20 percent of the market. “You don’t want to burn Intel,” Morales told EETimes. “If I am an AsusTek, I need to get processors for my other product lines from them.”

Forecasting netbook futures: even more fun than it looks

When analysts lock horns over Linux vs. Windows or Intel (x86) vs. everyone else, they are largely stomping over familiar territory; business firms have been discussing these questions for as long as the products in question have existed. Toss in a badly defined buzzword (Jon’s interview with Russ Daniels on the topic of cloud computing is absolutely excellent), a recession unparalleled in recent history, an entire new class of devices, and Intel’s “Atom Everywhere” strategy, and you’ve got an awful lot of muck inside the crystal ball.

Both predictions have merit. We’ve recently discussed how ARM+Linux makes sense for Nokia; NVIDIA has its own plans for Tegra, and there are other companies with ARM-based solutions in the works. Longer-term, even sub-netbook devices like Palm’s upcoming Pre could prove themselves as gamechangers—the eventual plan, after all, is to push richer, more fully featured experiences into smaller and smaller form factors (hence Intel’s entire vision of a MID).

The list of variables in play, however, is simply enormous. Dr. Castellano may be correct when questioning whether or not people will still buy netbooks as secondary systems in an economic recession, but he rather narrowly conceives of these devices as secondary computers, educational tools, and third-world primary systems. It’s not a good idea to box in the category so narrowly, particularly when speaking about device roadmaps 2-3 years away. True, Linux could capture additional market share from Microsoft in certain price-sensitive areas, but the success or failure of such a strategy will partly hinge on how strong of a price/performance ratio the software giant creates for netbooks with Windows 7. There’s also a need for better software tuning; it made headlines last year when MSI revealed that its Linux-equipped Wind netbooks were returned at a substantially higher rate than their Windows counterparts.

On the other hand, Atom’s ascendency is far from guaranteed. In addition to facing challenge from VIA and possibly AMD at a later date, strong netbook sales in the face of enduring economic hardship could actually encourage OEMs to thumb their collective noses at Intel. Morales is correct when he notes why Asustek might be unwilling to risk alienating Santa Clara, but OEMs prefer to offer differentiated product lines and must design such with an eye towards their own bottom-line figures. A resurgent AMD, meanwhile, could put an entirely new spin on the Atom vs. ARM equation—manufacturers who feel they can depend on the AMD/Globalfoundries team to supply them with microprocessors would presumably be more willing to experiment with additional non-Intel products.