Intel has been talking up its Atom processor since the beginning of the year, and it has clearly described its vision of a future where all of us are connected through a vast network of Mobile Internet Devices (MIDs). The 45nm Atom processors Intel has launched to date are the first arrivals in a product series the company intends to scale from handhelds to desktops, but current Atom deployment is still significantly bottlenecked.
The problem isn't yield or volume. Intel can currently build about 2,500 Atom processors on a single 300mm wafer, and yields are reportedly quite good. Santa Clara's testing facilities, however, are reportedly saturated at current production levels, and that's not a problem Intel can quickly resolve. There is always the option of converting test lines that are currently validating other CPU solutions over to Atom, but that simply transfers the backlog from one product family over to another. So far, as Infoworld reports, Intel has chosen to keep the majority of its testing capacity focused on products with significantly higher selling prices.
If the netbook market continues to explode, however, Intel may be forced to convert other testing equipment over to Atom or risk losing design wins. Manufacturers have practically fallen over themselves to introduce new netbooks this year; Asus might have broken the mold with the original EEE, but it's impossible to swing a dead cat more than six inches without running into a netbook design from MSI, Acer, Gigabyte, Asus, Dell, HP, Everex, and a half-a-dozen other companies.
Although HP is the only Tier 1 manufacturer to declare for VIA's Nano so far, a continued shortage of Atom chips could lead more manufacturers to consider VIA. Obviously, this isn't going to vault VIA to some position of dominance, but even modest gains would be big news for the smaller company. There are plenty of netbooks or MID devices Nano isn't capable of powering, thanks to its overall power consumption, but there are also plenty of higher-end systems where it could find a ready home.
As for AMD, at present, the company simply isn't focusing on the netbook market. We've heard rumors of an AMD-class netbook processor, and Dirk Meyer has promised to announce a new processor in November, but samples of any such chip shouldn't be expected to pop until mid-2009 at the earliest. Intel's validation constraints should be lifting by then, assuming the company's Vietnam facility comes online as scheduled.
If anyone benefits from Intel's constraint, I'd expect it to be VIA, but I also expect Santa Clara to keep a very close watch over Atom's growth. If the company feels that it is losing important sales due to production constraints, it might very well decide to take a (small) hit in another processor family in order to establish Atom's presence in as many first-generation netbooks as possible.
Thus far, we've discussed Atom's hardware support, but let's shift gears slightly and touch on software development. Linux developer Opened Hand has announced that it has been acquired by Intel and will henceforth be part of the Intel Open Source Technology Center. Opened Hand will focus on development of the Moblin Software Platform, which describes itself as "an open source community for sharing software technologies, ideas, projects, code, and applications to create an untethered computing experience across Mobile Internet Devices (MIDs), Netbooks, and embedded devices."
Moblin.org is explicitly focused on developing for Intel's Atom processor, and the "low power, low footprint, high performance, wireless, and graphics" capabilities Intel wants MIDs to offer in the future. Much has been made of Vista's inability to provide a compelling experience on netbook-class hardware, and Intel's purchase of Opened Hand (a company which a great deal of its own experience in Linux handheld development) is a rather clear statement of where Santa Clara thinks the netbook OS world may go.
It's been at least ten years since the technology world began talking about "Linux on the desktop." Now, despite the ongoing presence and popularity of Windows XP, "Linux on the netbook," may be a far more realistic goal.