Unnamed industry insiders claim that Samsung’s upcoming next-gen flagship, expected to be marketed as the Galaxy S10, will have a dedicated AI application processor, according to recent reports from Korea’s ETNews. Said to be built on the 7nm EUV process, the SoC will include at least two dedicated AI processing cores. That will bolster the chipset to a more competitive position when it comes to processing images, speech, and machine learning-related tasks compared to solutions on offer from other silicon manufacturers that are investing in the space. The chip and smartphone are expected to be revealed at the same time, at a launch event for the Samsung Galaxy S10 in the first half of next year.
Background: Although still speculative and lacking in deeper details, the new leak seems to add weight to prior leaks stemming from at least one former employee at the company. In effect, Samsung is thought to be working itself into a position that will allow it to directly compete with one of its biggest rivals in the Smartphone industry. Dedicated AI-enabled chips launched back in 2017 with Huawei subsidiary Hisilicon’s Kirin 970, followed up by the Kirin 980 just last month. The newest iteration offers a dual-core dedicated AI processor on top of a more standard SoC and is widely thought to be the best available to include a neural processing unit. Huawei is also the second largest manufacturer of Android handsets on the planet by nearly all accounts. That means that the top brand in the emerging AI chipset market is also the biggest threat to Samsung’s current position as the top OEM in the smartphone industry.
Simultaneously, Qualcomm is rumored to be revamping its own mobile chips to include dedicated neural processing units alongside the rebranding of its flagship SoC lineup. The first to include AI processing on separate cores is currently expected to be the Snapdragon 8155 and that is also rumored to be launching in early 2019, following its certification with the Bluetooth SIG in October. Samsung’s offering is expected to be marketed as part of its Exynos line of mobile and tablet processors, tentatively as the Exynos 9820.
Impact: If the rumors and speculation that have surfaced so far bear fruit, it’s not necessarily unlikely that the introduction of AI dedicated cores at the hardware level will spread well beyond flagship SoCs. Previously, when the market has seen new components introduced at the top end of the spectrum this quickly, that has been followed fairly quickly by those features appearing in budget-friendly or mid-range handsets across the board. Huawei has already started that trend with some of its own Android smartphones, specifically with those introduced under the Honor branding. Although that OEM’s chips have remained exclusive to its own devices, Samsung and Qualcomm still need to follow suit to stay relevant. For Qualcomm’s part, incorporating AI-dedicated cores will make its components more attractive to third-party smartphone manufacturers looking to compete with Huawei. For Samsung, the same holds true but with the added urgency created by its position in the overall smartphone market.