Samsung has tied itself to a new expedition that Samsung Semiconductor CEO Kye Hyun Kyung revealed. He took LinkedIn to announce the inauguration of Samsung Semiconductor AGI (Artificial General Intelligence) computing labs, which will be constructed in South Korea and the United States of America.
Under the leadership of former Google developer Dong Hyuk Woo, this lab purely lays down its focus to trailblazer an unprecedented semiconductor tailor-made to the computational demands of future artificial general intelligence.
Initially, the AGI Computing Lab will be immersed in developing chips optimized for large language models (LLMs), particularly delving into inference and service applications. Samsung Semiconductor is undertaking a nuanced reassessment of chip architecture, wide-ranging memory design, lightweight model optimization, high-speed interconnects, advanced packaging, and more, all geared towards substantially reducing the power consumption required for LLM operations.
Samsung’s plan involves a blazing-fast, iterative release of new chip designs from the AGI Computing Lab, assuring increasingly sturdy performance support for evolving models while minimizing power consumption and costs.
Last month, reports surfaced regarding Samsung’s venture into AI semiconductor logic chips and the establishment of a general artificial intelligence computing lab in Silicon Valley. This initiative, spearheaded by Dong Hyuk Woo, who previously made contributions to Google’s Gemini Nano AI model, marks a significant step for Samsung in the realm of AI technology.
Through the AGI computing laboratory, Samsung aims to tackle the Byzantine system-level challenges inherent in general artificial intelligence, ultimately offering accessible and sustainable solutions for the next wave of advanced AI/ML models.
