The battle between AI capable mobile chips heats up as four companies rise above the rest

The battle between AI capable mobile chips heats up as four companies rise above the rest
Artificial intelligence is one of the latest features you see mentioned in flagship phones’ ads and promotions. Still in its infancy, AI on smartphones currently handles tasks such as identifying objects you’re trying to take pictures of or optimizing the use of the processor to better fit your needs. While those aren’t exactly must have extras, the technology has great potential, which is why all the big players have a horse in the race: Apple with the A11 Bionic, HiSilicon with the Kirin 970, Qualcomm with its Snapdragon 845 and Samsung with the Exynos 9810. Those aren’t the only chips with AI, but they are currently the most capable, when it comes to mobile devices.

The combined efforts of these companies helped the application processors with AI market to grow almost three times, between the first quarters of 2017 and 2018. While such growth is unlikely to continue once chips with AI become the norm, it’s still an impressive leap forward in the span of 12 months.

One of the silicon manufacturers that’s unable to fully reap the benefits of this trend is MediaTek. The company lost market share five quarters in a row and it shipping numbers are currently at a four-year low. The company released its first fully capable AI chip, the Helio P60, earlier this year, but it’s mostly used in devices meant for the Asian market.

Later this year we’re expecting to see Apple’s A12 chip, Qualcomm’s Snapdragon 855 and the Kirin 980. All three are based on the 7nm architecture and should provide a significant boost to machine learning and AI capabilities.

Soon AI chips will start showing up in other household devices like tablets and smart speakers. Before you know it, you’ll be discussing cooking recipes with your oven.

source: DigiTimes



1. Atechguy0

Posts: 918; Member since: Aug 03, 2018

I don't know who does their research in this area, but clearly you guys are not doing a good job. Especially when there is standalone AI processors as well. You only mention AI accelerating co-processors, which are found in Apples SoCs, Huawei Kirin SoCs, Qualcomms SoCs, and Samsung's SoCs. But no mention of GPU based products, or standalone products like Google's Tensorflow chips, among others. Google's smartphone has a standalone AI processor called the Pixel Visual Core that is rated at 3 trillion operations per second.

2. walauweh

Posts: 9; Member since: Aug 08, 2018

apple bionic n kirin npu are also standalone ai chips. not denying how good is pixel visual core, but it only can do visual neural processing. kirin npu is cambricon 1m ai standalone chip, can do neural network process for audio n visual. qualcomm n samsung use mainly dsp, like mediatek p60 using cadence tensilica dsp. maybe u also need to do a bit of research. btw, all of them using heterogeneous ai api, meaning will work across all, cpu, gpu, dsp, npu.

Latest Stories

This copy is for your personal, non-commercial use only. You can order presentation-ready copies for distribution to your colleagues, clients or customers at or use the Reprints & Permissions tool that appears at the bottom of each web page. Visit for samples and additional information.
FCC OKs Cingular's purchase of AT&T Wireless