Qualcomm has enjoyed quite a smooth ride over the past few years, especially since they started integrating LTE into their SoC’s, which has had a huge effect on Qualcomm’s sales in LTE-based markets, but also in other markets because of the free publicity and popularity that Qualcomm’s chips received in the press.
While I think Qualcomm makes some of the very best chips in the mobile market, I was also starting to get worry that we might have another “Intel” in the mobile space – and not in a good way! Fortunately, with ARM’s open licensing model, it seems pretty easy for competitors to pop-up and disrupt Qualcomm or any company getting close to monopoly levels, and steal their thunder. That’s actually a good thing.
Qualcomm is already in a bit of trouble because of Apple, who released the first 64-bit chip this year, which is coming from a company that wasn’t even a chip maker until last year. Apparently, this move “hit them in the gut”, because I think Qualcomm, due to its huge market share in the mobile chip market, wanted to milk the 32-bit Krait core architecture for 3 years, instead of the usual 2 years, like everyone else in the industry.
Qualcomm probably thought they pretty much own the market, so there was no need to move to ARMv8 that quickly. Instead, they could just re-use the same architecture that was launched in early 2012, and optimize it further, which would be much cheaper for them (i.e. higher profits), than to switch to ARM on time (not early, just on time).
But Apple may not even be Qualcomm’s biggest worry, at least in the longer term. Low-cost chip suppliers like Broadcom (Raspberry Pi maker), Mediatek and Spreadtrum are starting to see huge growth in the Asian markets, as well as in other LTE markets, too, because they are now coming out with their own low-cost LTE-integrated chips, and OEMs are starting to prefer them over Qualcomm’s own chips.
Qualcomm doesn’t seem care a lot about the low-cost competition, at least going by how they think about releasing competing chips. Unlike at the high-end, where they design the best CPU they can possibly make in that category, and then use lower-clocked versions for mid-range, at the low-end they just take ARM’s own stock Cortex A5/A7/A53 cores, and then add their own GPU and LTE to them.
Those chips can also benefit from their “Snapdragon” branding (even though they aren’t really “Snapdragon” chips). They just don’t seem to care enough about the low-end markets to get engineers to make their custom low-end CPUs targeted at them. It’s more profitable for them to just use ARM’s stock low-end CPUs (which do seem pretty great for their market, I have to say), so they stick with that.
In the end, we all benefit from more competition, so I actually don’t mind this happening at all.