Samsung began to build its own ultra-large AI system

The announcement implies that Samsung has already devoted considerable resources to its proprietary ultra-large AI system, leveraging it to catalyze the evolution of artificial intelligence technology applications.

This ambitious endeavor is spearheaded by the Samsung Research Institute and launched in earnest in the early days of June, with a substantial allocation of GPUs for model training. The aspiration is to conclude the training of the inaugural ultra-large AI system by the end of July.

The relevant technology will initially be utilized internally, with the potential expansion to consumer market products remaining uncertain for now. Early-stage functions encompass document summary extraction, language translation, and assistance with software development, thereby bolstering internal operational efficiency. This could, for example, curtail the time necessary for software development or expedite the semiconductor research and development cycle.

Hitherto, Samsung has sternly prohibited its internal workforce from employing AI tools provided by entities such as OpenAI, Microsoft, or Google, to forestall the leakage of confidential information. Furthermore, Samsung’s chairman, Lee Jae-yong, is a staunch advocate for swiftly investing in proprietary AI technology research to diminish reliance on external technological resources.