In a recent post, I wondered aloud whether Apple would strike out on its own, devising its own components around Intel CPUs in order to acquire some feature or performance advantage at the expense of reliance on commodity component parts. I thought it was the kind of thing that could deliver the margins decrease with which Apple guided analyst expectations in its last quarterly conference call.
This article says, "no" and predicts Apple will use nVidia's hunger to sell chipsets as a mechanism to fob off on third parties its interest in differentiating products with new tech, without risking generalized compatibility or undertaking to compete with every chipset vendor on the planet to make a commodity it should be able to get at bid.
This leaves open the question what Apple will do to decrease its margins, and what product transition would have the effect of driving up costs.
One alternative: by beginning -- in its portables business -- to beef up GPUs, Apple could begin investing in costlier graphics subsystems, which would certainly create an opportunity to drive up costs. GPUs can be very pricey, after all. The reason this can impact margins is that (a) Apple will be competing with notebooks that have typical, weak GPUs and non-innovative chipsets, and it's unclear that everyone will pay for quality. Just increasing the cost to match the component increases -- that is, making cost back to preserve per-unit profits -- would result in a margins decrease even if both units and aggregate profits grew.
However, since Apple can't project demand and won't want to get caught predicting a landslide of sales that doesn't materialize, Apple will predict conservative units and sales while it's predicting lower margins, which of course yields a dismal picture of the future.
The technology in Apple's next operating system, to enable applications to leverage GPUs to accelerate run-of-the-mill calculations, needs a hardware basis now to create demand for the next software upgrade. If Apple lays a foundation for software demand at the cost of a little margin now, it may reap increased margins as performance-drive OS demand creates high-margin software sales next year or so. The questions, of course, include (a) at what price will Apple be able to sell its GPU-heavy notebooks, (b) will Apple's GPU-heavy notebooks increase sales notably (will it encourage gamer purchases? will gamers buy only big high-octane towers?), and (c) how big will demand be for an OS which advertises speed, performance, and efficiency as its principal upgrades?
Given that (c) has little parallel outside Apple (MSFT upgrades tend to require hardware upgrades, not feel like hardware upgrades), and Apple's comparables typically involved high-profile applications or features (e.g., Time Machine), it could be hard to tell. Developers who write Snow Leopard enabled (or exclusive) apps might tend to pull users along -- as they did with Tiger and Leopard -- but will Snow Leopard offer developers features that lead to Snow Leopard exclusive applications, or just applications that happen to perform better under Snow Leopard?