Yeah but really they never had a chance in the mobile market. No way. But they owned the PC market and blew it by improving nothing from 2010 through 2018. That's much less forgivable.
> Yeah but really they never had a chance in the mobile market.
You're saying that if Intel had used their leading edge semiconductor fabs to make ARM CPUs that were exceptionally good, they would not have stood a chance in mobile? I disagree. Qualcomm would have faced serious challenges keeping up with these hypothetical Intel ARM processors back in the early 2010s. They still might, but Intel hasn't even tried.
Intel was just too proud to do something like that. They wanted x86 to become the smartphone standard, which it obviously didn't... and now they're suffering defeat on all fronts.
It would also require a different type of design flexibility.
The mainstream Intel CPU line is relatively limited in variations. Yeah, we have different core counts and presence of SMT or iGPU, but you can get there with basically a single die design, or a small family of them, and blowing some fuses after binning.
Mobile SoCs tended to be a lot more bespoke. The same basic CPU block might need to e paired with different modems, and physical space and production costs probably don't accommodate doing the "low-end by fusing off bits of big higher-end chips" model.
As I understand it, Intel wasn't fond of that sort of product diversity-- the Atom they developed for mobile devices was much more a take-it-or-leave-it proposition.
I also suspect the concept of being an external single source was unappetizing for buyers. Samsung has the choice of buying Snapdragons or making their own Exynos parts; low end manufacturers can cross-shop MediaTek or Allwinner. If you're too reliant on a uniquely Intel design, what happens if they have a production kink and you have a million phones awaiting processors?
Mobile is dominated by cost. Qualcomm, Samsung, AMD have all tried their hand at good ARM processors but the perf gains and costs have never been good enough to sell mass market devices. Apple is the only one remaining, and they spend a lot per chip... it would be about as expensive as an Intel chip, if not more expensive, if sold separately.
> They wanted x86 to become the smartphone standard
Or that the future was Netbooks with Atom CPUs. But yes, they bet very wrong and didn't hedge that bet. I think Intel was even an ARM licensee at the time, and I'm sure ARM would have been happy to take more money from them.