The alleged deal tweaks the nose of Intel, which had pushed its Atom processor at a range of embedded devices. But Dan Dobberpuhl, who founded PA Semi, also helped design the StrongARM chip, which Digital Equipment Corp. helped design before the technology was sold to Intel in 1997. Intel then tried to use StrongARM to batter its way into the emerging phone market, and largely failed.

Now, the PA Semi chip could become the engine behind a next-generation iPhone and complementary devices. Why would Apple, a traditional software powerhouse, want to own a chip company? That's a little less clear. Apple still has the clout to demand custom designs, although not, apparently, from Intel. I suspect that the company could have maintained a traditional contract relationship with PA or its competitors and done well.

Here's the thing, though: in the electronics industry, there's either a build-or-buy choice to be made. Buying a processor company means Apple has ultimate control over the supply and direction of PA Semi. But it also locks itself in to using a specific embedded microprocessor, while there's nothing preventing the next new hot startup from suddenly appearing and selling its own processors to Apple's competitors. Remember, Apple has a well-deserved reputation for building in the latest hardware technology into its products in new and interesting ways. But Forbes reports -- and it's right, generally -- that it will take about a year for a custom chip to come to market. I'd suspect that Apple could use an existing PA Semi chip well before that.

I would think that all would be revealed Wednesday afternoon, when Apple reports its first-quarter earnings. And you know what? With $9.1 billion in cash on hand at the end of 2007, Apple can afford to place (and possibly lose) a $278 million bet.

No comment from PA Semi and Apple yet...