Galaxy S26’s Exynos 2600 Could Make On-Device AI Smarter With Nota AI

Samsung has pushed harder into on-device AI capabilities with its latest Exynos 2600 chip. The chip leverages AI model optimization technology from Nota AI to enable more efficient AI processing directly on the device. If the chip powers the upcoming Galaxy S26 and S26+, as rumored, it could bring more advanced offline AI experiences for flagship users.

Samsung teams up with Nota AI for on-device AI optimization in Exynos 2600

Nota AI provides technology that shrinks and optimizes AI models, allowing them to run efficiently on devices. Thanks to its Netspresso, it can reduce model size by up to 90% while maintaining accuracy. This makes it possible to run large-scale AI tasks directly on smartphones without depending on cloud servers.

According to a report from ETNews, Nota AI has signed a supply agreement with Samsung to provide its AI optimization technology for Exynos 2600. This collaboration will also allow the Korean firm to improve its Exynos AI Studio, a toolchain that helps developers optimize and deploy AI models on a chip. It is worth mentioning that Nota AI also provided its optimization technology for the Exynos 2400 and Exynos 2500.

“The collaboration that has continued since the Exynos 2400 proves that Nota technology is creating real value when combined with Samsung hardware,” said Chae Myung-soo, CEO of Nota. “Through continued collaboration with global partners and technology commercialization, we will firmly establish Nota as an essential technology leading the on-device AI era.”

Meanwhile, Exynos 2600 also brings substantial hardware improvements. Based on Samsung’s 2nm Gate-All-Around (GAA) process, the chip offers 39% better performance than its predecessor. It features a 10-core CPU based on the Arm v9.3 architecture and Xclipse 960 GPU. Its new neural processing unit (NPU) delivers 113% improvement in AI performance over its predecessor. Thanks to Heat Path Block (HPB) technology, the chip can manage heat during demanding tasks such as gaming or AI processing.