TapTechNews on June 18th, it was reported that the corporate governance report recently released by Samsung Electronics showed that the board of directors' management committee of Samsung Electronics passed the proposal of investing in GPU on March 19th this year.
The management committee of the board of directors of Samsung Electronics is composed of the core management team of the enterprise and is responsible for the review of major matters of Samsung Electronics. The management committee has been publicly disclosing the specific agenda since 2012.
According to the South Korean media Bloter, the previously announced agendas mainly focused on factory construction and equipment investment in the storage and logic fields. Recently, with the surge in the demand for AI semiconductors, the investment decisions in advanced packaging have increased sharply, but before March 19th, there has never been an investment decision for a single system semiconductor.
TapTechNews note: According to the dictionary on the official website of Samsung Semiconductor, a system semiconductor refers to a semiconductor with logical operation and control functions, covering microdevices (such as MCUs and DSPs), analog chips, logic chips, and optical semiconductors.
Due to the rather ambiguous expression in the corporate governance report, it is currently difficult to determine the specific direction of the investing in GPU resolution of Samsung Electronics. South Korean media believe that there are two main possibilities:
Samsung Electronics not only has experience in contract manufacturing GPUs, but also recently cooperated with AMD to transplant RDNA GPUs to Exynos processors, and it can also provide HBM memory. Investing in the GPU field can enhance the profitability of these businesses.
The HPC center in Hwaseong of Samsung Electronics was completed in April this year, and a large number of GPUs will be introduced in the future to meet the needs of semiconductor production; Samsung Electronics will also build a digital twin chip factory, which also requires a considerable amount of GPU computing power resources.