Qualcomm is putting dedicated AI hardware directly inside the mobile graphics pipeline of its next flagship Snapdragon processor. A new Adreno GPU redesign features three dedicated Qualcomm’s Matrix Cores designed to run AI locally during rendering. By moving these workloads straight to the GPU slices instead of shuttling data back and forth to main system memory, Qualcomm cuts latency, reduces thermal throttling, and lowers overall battery consumption during demanding mobile gaming sessions.
Adreno Neural Fusion and upscaling tech
These new dedicated cores power a feature suite dubbed Adreno Neural Fusion. They act as Qualcomm’s mobile answer to PC upscaling solutions like NVIDIA DLSS. By running AI super-resolution and frame generation algorithms right where graphics processing happens, the system eliminates common visual artifacts like ghosting and shimmering while maintaining much smoother frame rates compared to older features like Snapdragon Game Super Resolution.
Power efficiency remains a huge selling point for mobile gamers. Qualcomm promises up to a 40% reduction in power consumption when Adreno Neural Fusion is active compared to previous upscaling methods. Game developers won’t need to build custom integrations from scratch either, as both Unreal Engine and Unity have already built native support for the Neural Fusion stack directly into their graphics engines.
Higher clock speeds, hardware specs, and Summit details
Under the hood, Qualcomm cleared up earlier rumors of a six-slice design. The firm confirmed a refined three-slice architecture where shader cores and hardware blocks are divided between slices for granular power balancing. The new Adreno GPU runs at a 1.45GHz clock speed. This is a noticeable jump from the 1.2GHz clock found on the Snapdragon 8 Elite Gen 5 GPU.
To keep data pipelines flowing smoothly, Qualcomm retained its 18MB Adreno High Performance Memory (HPM) cache. By keeping frame buffers, compute workloads, and tile-based rendering local to the graphics subsystem, the chip achieves a claimed 12% boost in overall efficiency while feeding the CPU. This GPU pairs alongside a 5GHz Oryon CPU supported by FlexCache architecture, an X105 modem, and FastConnect 8800 connectivity. Full details regarding the NPU and official branding will drop at Qualcomm’s Snapdragon Summit in Maui, Hawaii. The event will take place from September 22 to 24.
The post Qualcomm Brings AI Matrix Cores Directly to New Adreno GPUs for DLSS-Style Upscaling and Frame Gen appeared first on Android Headlines.
​Â
