12 April 2019
A new report from an analyzing firm deduces that three out of four phones will have a dedicated AI chip by 2020. If this happens, the phones would hold improved brains in just 3 years from now. Better performance and less dependency on the internet is what the current companies thrive for. However, this would take some more time. But, seeing the smartphones or the devices we use getting improved brains is no lesser than any later stage arriving after it. This is the start of the evolvement of some intelligent tech and a dedicated AI chip would let devices accomplish some assigned tasks on their own without depending on their owners. It will only improve like human brains over the period of time by learning the habits of its owners.
The report from Counterpoint Research puts stress on the number of sales of AI smartphones that would increase by 2022. To be precise, the report claims the sale would increase to 1,250M units in 2022 from 190M in 2018. This is a huge growth that would get its catalyst from the inclusion of AI. This means AI chips would help in unlocking some features aren’t possible to execute now and smartphone manufacturers would utilize this untapped power so that buyers would attract.
Note that some AI features are already utilizing the device-based processing without any dedicated AI chip, for instance, the voice assistants. Now, a dedicated AI chip which would become an ultimate reservoir of power for future smartphones so as to unbolt some capabilities which don’t seem possible now. What we could imagine for now is the faster processing of commands on-device performance.
At present, AI is being used mostly for camera-related applications (facial recognition, image enhancement, and more), but according to the report, this would change as we would get to see "AI being used extensively for a raft of other applications over the next two years."
Huawei and Apple are already the OEMs which have included dedicated AI processors in their system on chips- Kirin 970 and A11 chips. Other companies have started following suit as well. For instance, Qualcomm is offering an AI Tensor Accelerator in the Snapdragon 855’s Hexagon DSP for the first time.
The report concludes that SoC vendors are competing to improve the machine learning (ML) capabilities of their chips by integrating dedicated AI processing cores into their designs.