- The Exynos-powered Galaxy S26 might feature better on-device AI with native image generation capabilities.
- Samsung is reportedly collaborating with another company to bring a lightweight version of Stable Diffusion for on-device image generation.
- The feature is expected to be called EdgeFusion and could generate 512-pixel images in “a few seconds.”
AI has now mutated into an integral part of our devices. But brands still use a hybrid AI arrangement, where only part of the processing occurs locally, with most of the lifting done by AI models in cloud infrastructure. On-device AI on phones, for instance, has largely been limited to processing text, while richer media, such as images, are processed online. As expected, the latter approach raises privacy concerns, but Samsung could take a step toward addressing them by bringing on-device image generation to the upcoming Galaxy S26.
According to the information received by the Korean news outlet Seoul Economic Daily, Samsung’s Exynos 2600 chipset will be capable of running text-to-image generation AI models without the internet on the Galaxy S26. The feature is likely to be called EdgeFusion (translated from Korean) and is based on Stable Diffusion image-generation models. The feature is developed in partnership with the Korean AI company Nota AI, which specializes in optimizing models for lightweight deployment.
