Samsung is now looking forward to running AI directly on the device, which helps to work faster, smarter, and more privately without relying on cloud servers.
Samsung is one of the leading brands in terms of AI technology. It has its own AI suite called Galaxy AI, which includes several AI tools, but most of those are based on cloud servers. Meanwhile, the company has now shared new information regarding the development of local AI. It’s highly anticipated that the upcoming Galaxy S26 series will be the first generation to fully inherit these optimizations.
Despite being integrated with powerful chips, large RAM, a bigger Vapor Chamber, and enough storage, the current generation flagship phones are not completely compatible with running complex AI models. Since larger AI models demand to perform billions of calculations directly on the device, this causes heat, drains battery, and possibly reduces performance.
To deal with all of these cautions, Samsung shifts its focus to model compression techniques that efficiently reduce the size of AI models while maintaining high accuracy. The company explained that in this process, quantization significantly simplifies complex calculations into more optimal integer forms, similar to compressing high-resolution images while maintaining almost unchanged quality.
Regarding this, Dr. MyungJoo Ham, an expert at the AI Center at Samsung Research, stated, “AI will learn better over time on the device and adapt to each user’s environment,” and added, “Our future is natural, deeply personalized services that still ensure privacy.”
