Samsung establishes Artificial General Intelligence Computing Lab led by former Google TPU developer in Silicon Valley

When it comes to cutting-edge AI developments, Samsung has remained mostly on the sidelines, since it just started adding more useful AI functions to its flagship smartphones and previously only focused on HBM DRAM to complement AI processors. However, the South Korean giant now appears to gear up and secure a more prominent position especially in the AI processing sector, which is dominated by Nvidia’s GPUs and to a lesser degree AMD’s models. Current GPUs and NPUs are becoming quite proficient at generative AI tasks, but Samsung is planning to develop Artificial General Intelligence processors that better emulate the intricate ways of the human mind. To this end, Samsung is establishing a separate research organization known as the AGI Computing Lab on US soil in Silicon Valley, as reported by MK Korea.

Samsung’s AGI Computing Lab will be led by former Google Tensor Processing Unit (TPU) developer Dr. Woo Dong-hyuk. The Tensor Processing Unit (not to be confused with the Tensor processor) is among the first large-scale neural networks introduced in 2016. Recruitment for microarchitecture key positions are now in full force and Samsung is looking to secure a sizable portion of the $100 billion AGI investment fund prepared by Softbank.

Meanwhile, Samsung’s foundries are also getting involved in advanced AI developments, as the company recently managed to land an important contract with the biggest AI developer in Japan and may continue to attract big players lie Microsoft, IBM, Arm and even Nvidia, which can soon become Samsung's direct competitor in the AGi processor sector.

Buy the Galaxy S24 smartphone with Galaxy AI features on Amazon