Qualcomm will add artificial intelligence cores called Adreno Matrix Cores to the GPUs of its upcoming flagship Snapdragon platforms for the first time. These units will sit at the center of Adreno Neural Fusion, combining neural network processing, AI-powered super resolution and frame generation within the graphics pipeline. The GPU will thus be able to render games at a lower resolution and produce higher-resolution images, while reducing flicker, ghosting and blurring. The new GPU aims to reduce the need to send data to the NPU and lower energy consumption. According to 9to5Google, each section of the three-part GPU will reach 1.45 GHz; the Snapdragon 8 Elite Gen 5 GPU reaches 1.2 GHz. While 18 MB of Adreno HPM will be retained, data flow will be improved with Matrix Cores rather than through increased capacity. This memory keeps tile-based rendering data, frame buffers and computation data close to the GPU, reducing the need to access system memory.
Why it matters
This architectural change aims to rebalance the trade-off between graphics performance and energy use by enabling AI processing to be carried out closer to the graphics pipeline. For gamers, its impact will be assessed based on how the computational workload is distributed while producing high-resolution images, and the extent to which flicker, ghosting or blurring issues in the image are reduced. Reduced data flow between the NPU and GPU could change how system resources are used, particularly in AI-powered imaging features. However, the available information does not explain which games and applications will use this design on actual devices, how much performance or energy savings Matrix Cores will provide in practice, or whether additional adaptation will be required for developers.
Background
GPU is not a new name in the FikirPilot archive: We have published 6 articles mentioning this name in the last 90 days; the latest was dated September 4, 2026.