Monday, September 05, 2011

Ultrabook: Intel's $300 million plan to beat Apple at its own game

ars has an interesting article, Ultrabook: Intel's $300 million plan to beat Apple at its own game. It's about the difficulties the PC industry is having in replicating Apple's MacBook Air. Intel has put up quite a bounty to help. After you get through a page of complaints about Apple's laptop keyboards, particularly for UK users, you get to some good stuff in the article. Having just helped someone buy a Dell laptop I can attest, the buying experience sucked. There were too many models that I couldn't distinguish between and too many pointless choices to make. Here's some of my favorite parts of the article, though there's more along these lines.

"In theory, it should be easy for the PC OEMs to match Apple. Intel insists that the manufacturing cost for 11-13", 18mm thick Ultrabooks should be between $493 and $710, while 14-17," 21 mm thick ones should cost between $475 and $650. If Intel's prices are accurate, there should be plenty of room to undercut Apple and still achieve reasonable margins. Sell machines for $949 and take home 25 percent gross? That's much healthier than Lenovo's current 12.5 percent average gross margin, for example.

Yet the PC OEMs say there's no way they can bring a product to market for less than Apple; even with slimmer PC margins, they can't get below $999. Instead, they're arguing that they need a 50 percent discount from Intel to beat Apple on pricing. Intel is willing to offer a 20 percent discount on its processors and marketing assistance, but it refuses to take the substantial margin hit that a 50 percent cut would require. Even if the chip giant did offer such a discount, it's difficult to see how it would help; the cheaper processors would just pad Apple's margins, or allow the Cupertino-based company to cut prices itself. Nobody wins in a race to the bottom.

The problem is that the PC industry, particularly the large OEMs, just aren't set up to produce this kind of machine. The PC industry is built around an idea of almost infinite variation: different Wi-Fi adaptors, different Ethernet chipsets, different GPUs, different USB3 controllers. This variety is then reflected in the systems available from manufacturers—and more importantly, it's reflected in the way the systems are actually built.

Consider Lenovo. Lenovo offers a range of different Wi-Fi adaptors in many of its systems. Instead of designing several different motherboards, each with a different integrated adaptor, it puts the adaptors themselves onto daughtercards and plugs them into a socket on the motherboard. The upside is that Lenovo can offer a lot of diversity, and the daughter cards can be standard Mini-PCIe components that anyone can use. The downside is that Lenovo has boards that are less integrated—hence larger—with more components and more complex manufacturing. Lenovo also has to buy smaller numbers of more adaptors than it would if it just picked one version and standardized. It also means that people buying Lenovo systems have to make dumb choices on the website."

No comments: