December 30, 2025 – According to a blog post published by South Korean media outlet ETNews today, Samsung’s upcoming Exynos 2600 chipset is set to incorporate an AI model optimization solution from Nota, a South Korean AI technology firm. This integration will enable a remarkable reduction of over 90% in the size of AI models while maintaining their accuracy.
As per the blog post, Nota, a company specializing in AI optimization technology, has revealed that it has inked a technology supply agreement with Samsung Electronics. Under this deal, Nota will supply its AI model optimization solution for Samsung’s next – generation mobile processor, the Exynos 2600. This marks the second time that Nota has secured a flagship chip order from Samsung Electronics, following its previous collaboration on the Exynos 2500.
The core technology provided by Nota this time is centered around its self – developed “NetsPresso” platform. This platform is capable of efficiently optimizing and deploying AI models across various hardware environments.
NetsPresso stands out for its powerful model compression capabilities. It can shrink the size of AI models by more than 90% without compromising the model’s inference accuracy. The primary objective of this collaboration is to enable the smooth operation of large – scale generative AI models on the Exynos 2600. With Nota’s optimization technology, mobile devices will be able to handle complex generative AI tasks locally, without relying on an internet connection.
In addition, Nota will play a significant role in the development of the next – generation version of Samsung Electronics’ AI model optimization toolchain, “Exynos AI Studio”. By automating the optimization pipeline, this tool aims to significantly streamline the process for developers to deploy the latest AI models on the Exynos platform, reducing development complexity and accelerating time – to – market.
