Intel (Nasdaq: INTC) has an earworm. Like the insidious piano intro to Journey's "Don't Stop Believing," the idea of stretching a bit to create new markets appears to be stuck in Intel's head, and we keep hearing the same song over and over again.
Weeks ago, the semiconductor superpower announced a bold redesign of its Atom line, aimed squarely at the mobile market. The doctor is out until system builders have had a chance to get familiar with that product, while QUALCOMM (Nasdaq: QCOM), NVIDIA (Nasdaq: NVDA) and other ARM Holdings (Nasdaq: ARMH) based chips try to lure hardware makers away from Intel by offering power-sipping but less-powerful processors across a range of mobile devices.
This time, Intel is casting bedroom eyes at the supercomputing sector. By packing as many as 50 simple execution cores onto a single chip, Intel aims to give researchers and engineers the number-crunching power they need to solve the hardest computing problems they can think up. Born from the fizzled Larrabee graphics processor project and a few other products that never quite made it to market, the Knight's Corner processor will go head to head against the best high performance graphics cards and Opteron processors offered by AMD (NYSE: AMD) and NVIDIA's Tesla chips.
However, with Knight's Corner, Intel will also cannibalize its own product offerings. The aging Itanium processor was built for exactly the kind of workloads that would benefit from Knight's Corner's massively parallel programming model. In the supercomputing space, Intel is mostly replacing itself.
It has to be done eventually, but don't expect Knight's Corner to improve any part of Intel's business, it should only hold the line. Intel is still a fine stock, but for other reasons.
Could someone please silence that piano?
