Samsung Electronics has teamed up with South Korean-based AI optimization specialist Nota AI to supercharge the on-device AI performance of its next-generation Exynos 2500 processor, by embedding Nota’s model compression tech directly into Samsung’s Exynos AI Studio toolchain.
The partnership targets on-device generative AI, as it will allow for bigger language models to run smoother and drain less battery on Samsung mobile devices, especially the ones powered by the 3nm mobile processor. Additionally, developers can also now optimize complex AI workloads for the Exynos 2500’s beefed-up NPU without cloud limitations.
Samsung positions this collaboration as a game-changer for mobile AI, otherwise known as On-device AI, as it lets apps handle complex tasks like real-time image generation or language processing right on the device.
“The integration of Nota AI’s optimization technology into Exynos AI Studio has enhanced the efficiency of on-device AI model development compared to previous generations,” Chulmin Jo, Vice President of Mobile AP S/W Development at Samsung Electronics, said in a statement. “We are pleased with the strong results from this collaboration and look forward to continued advancements together.”
Nota AI’s Optimization Technology
Nota AI is a key player in the field of AI model optimization, with its flagship platform, NetsPresso, at the forefront of its product offerings. NetsPresso employs a four-step process that involves the training, compression, compilation, and benchmarking of tailoring AI models for devices with limited resources.
This hardware-aware optimization is crucial for deploying complex AI functionalities, from computer vision to generative AI on smartphones. Nota AI has also collaborated with other industry leaders such as Qualcomm, to streamline the development and deployment of AI and Large Language Models (LLMs) on a variety of devices.
Why Mobile AI Hits Different Now With An Upgrade to Exynos 2500
On-device AI isn’t new, but making it scale on everyday phones changes everything, as the former cloud reliance meant latency, privacy risks, and dead batteries during spotty connections.
But now, with Nota AI’s tech, this translates to smarter cameras that can edit photos on the fly and games with AI-driven NPCs that customers can easily access.
For Samsung, it sees this as a way of popularizing generative AI on mobiles, especially as rivals like Google and Apple push similar boundaries.
For Nota, which is fresh off a KOSDAQ IPO with backers like Samsung Ventures and NAVER, this integration signifies its vision of a global expansion as it validates their AI-optimization focus. Myungsu Chae, CEO of Nota AI, highlighted the significance of this partnership, stating that it represents a “breakthrough case in bringing their AI optimization technology to globally commercialized consumer devices.”
By integrating Nota AI’s expertise, Samsung also aims to deliver a substantial boost in optimization efficiency and performance, which paves the way for advanced generative AI applications to run seamlessly on the Exynos 2500 chipset.
