Samsung Prepares Rp1.247 Trillion Investment to Dominate AI Chip Market, Surpassing TSMC

Samsung Electronics plans to allocate more than 110 trillion won, or Rp1.247 trillion, for expanding capacity and research into artificial intelligence (AI) chips by 2026. This allocation represents the largest annual capital commitment in the company’s history, aimed at securing a leading position amid the explosion in global technology demand. The company’s disclosure shows that this investment value increases by 21.7% compared to last year’s realised expenditure of Rp1.024 trillion. This figure also marks the first time the South Korean technology giant’s annual spending exceeds the 100 trillion won threshold, or about Rp1.134 trillion. Samsung’s expansion strategy surpasses the capital expenditure estimates of Taiwan Semiconductor Manufacturing Co. (TSMC), which has set aside around US$50 billion for similar purposes. This decision signals a shift in the company’s focus to respond to AI-driven demand. Samsung’s Co-Chief Executive Officer (CEO) Jun Young-hyun stated that the emergence of agentic AI has triggered a massive surge in orders for server-class memory and storage. The company is now prioritising the development of next-generation AI chips and more advanced foundry processes. “The emergence of agentic AI has triggered an explosive surge in orders, not only for high-bandwidth memory (HBM) but also for server-class storage,” Jun said, as quoted from The Business Times on Friday (20/3/2026). These efforts aim to reclaim competitive advantage from SK Hynix Inc., which currently dominates the supply of high-bandwidth memory (HBM) for Nvidia Corp. Samsung targets a dominant position through the development of more advanced AI chip technology. On the other hand, Samsung is also planning to supply the latest generation HBM4 chips to OpenAI. These chips will be integrated into OpenAI’s first internal AI processor, which is being developed in collaboration with Broadcom. Samsung has signed a letter of intent to supply memory for OpenAI’s data centres in the “Stargate” project. The delivery target is 800 million Gb of 12-layer HBM4 chips, scheduled to begin in the second half of this year. In addition to OpenAI, Samsung is strengthening its strategic partnership with AMD through the signing of a memorandum of understanding (MoU). In this collaboration, Samsung will serve as the primary supplier of HBM4 chips for AMD’s future AI graphics processing units. To support mass production targets, the company is improving efficiency at the P4 Pyeongtaek plant and installing equipment on the P5 production line. Construction of a new manufacturing facility is also underway in the Yongin chip cluster, Gyeonggi Province. Outside South Korea, Samsung is building a new foundry plant in Taylor, Texas, United States. This manufacturing facility is targeted to begin operations by the end of 2026 to meet high demand in the North American market. This capacity expansion is expected to help address the global traditional memory deficit that has begun to disrupt various industries. The industry’s shift to high-end memory has previously caused chip shortages for sectors ranging from automotive to smartphones. Samsung’s management has also revealed plans to pursue mergers and acquisitions in the robotics, medical technology, and automotive electronics sectors. This diversification is intended to strengthen the business portfolio beyond the core semiconductor sector.