why intel deserves its current problems

The news is out that Intel is abandoning the mobile segment and laying off 12,000 employees world-wide. The primary issues are ARM and it’s deeply entrenched hold on the mobile marketplace combined with a decline in the overall PC market, due in no small part to everyone moving to ARM-powered mobile computers called smartphones. Intel got into this position quite frankly because of its arrogant X86-only worldview.

Intel used to make ARM-based processors nearly 20 years ago. Intel got into ARM-based manufacturing when it purchased Digital Equipment Corporation’s (DEC) StrongARM division in 1997 as part of a lawsuit settlement. Intel manufactured StrongARM until around 2000, when it introduced XScale, a new at-that-time ARM-based series of chips. XScale production continued until around June of 2006 when the XScale line was sold to Marvel. Once that sale was consummated Intel turned back to its X86 chips and attempted to manufacture low-powered versions that would supposedly compete against ARM in the mobile processing market. Intel, unless they paid (bribed) a vendor to do so, never won any significant design wins.

This situation continued for nearly a decade, with Intel spending billions in R&D and marketing, trying to push into a market its prior leadership never took seriously until way too late. The biggest killer of the Atom SoCs has been Apple, with its A-series of ARM-based chips, especially A8 and now A9. Apple has pushed performance-wise to within a very short distance of Intel’s low-end i3 entry level X86 chips, and blown the Atom series of embedded chips out of the water in the process. It was Apple after all that helped develop ARM back in the early 1990s for use in the Newton hand-held device from that era. Apple got back into custom ARM design in a big way when it purchased P.A. Semi in 2008 and Intrinsity in 2010. Both those strategic purchases gave Apple a lean, mean design team that helped create a power stingy, yet computationally powerful family of processors for its line of products. While it’s doubtful that ARM will replace Intel in Macs, ARM is already in the majority of Apple products that make the lion’s share of Apple profits. And none of those processor profits flows back to Intel because there are no Intel processors in those mobile Apple products.

Intel could have been in a better position technologically and financially if they’d kept the XScale line. But they didn’t own ARM IP the way they owned and controlled X86 IP. Their focus on control coupled with short-term profitability blinded them to the long-term trends in mobile computing that were obvious even in the mid-2000s; away from anything X86 and towards ARM in general. Add in the Intel hubris about X86 and you have a nasty combination that’ll be studied in business schools for many years to come. Intel won’t go out of business any time soon, and they may yet evolve into something smaller that provides better long-term survivability, but their days of being bullying Chipzilla to the rest of the semiconductor world is over.

Long live ARM.