DEEPX has signed a development agreement with Samsung Foundry and design partner Gaonchips to build its next-generation DX-M2 processor on Samsung’s 2nm Gate-All-Around (GAA) process technology.
With this collaboration, DEEPX becomes the first commercial customer of Samsung Foundry’s advanced 2nm process, marking a major milestone in the global semiconductor and AI ecosystem. The DX-M2, designed for ultra-low-power generative AI inference, is expected to begin prototype production in the first half of 2026, with mass production slated for 2027.
The move is seen as strategic for DEEPX, which aims to overcome the computational and power constraints of existing process nodes. The 2nm process offers nearly double the power efficiency of the current 5nm technology used in DEEPX’s DX-M1 chips, making it a fundamental enabler for real-time, on-device generative AI applications.
The DX-M2 chip is capable of running 20-billion-parameter AI models in real time at 20–30 tokens per second, all while consuming under 5 watts of power—a breakthrough compared to current industry standards. This efficiency leap enables robots, smart appliances, and laptops to run advanced AI models independently, without reliance on energy-intensive data centers.
DEEPX expects the project to significantly boost South Korea’s high-value semiconductor industry, while setting a new benchmark for practical, scalable on-device AI.
“DX-M2 is the core platform that will usher in the mass adoption and industrialization of generative AI,” said Lokwon Kim, CEO of DEEPX. “By combining 2nm process efficiency with our AI architecture, we aim to eliminate the cost and energy barriers of today’s AI and create a world where anyone can benefit freely and sustainably from this technology.”
(This content is sourced from a syndicated feed. The Japan India Manufacturing Journal website assumes no responsibility or liability for its accuracy, completeness, or content.)
COMMENTS