Samsung to record $100 mn global revenue from chip-packaging products

Samsung’s revenue estimate: Samsung Electronics is eyeing to generate $100 million or more in revenue from its upcoming batch of advanced chip-packaging products this year, as per the company’s co-CEO Kye-Hyun Kyung.

Last year, Samsung established advanced chip packaging as a dedicated business unit, with Kyung expressing confidence that the outcomes of Samsung's investments will materialise in the second half of this year. 

These statements were made during Samsung's annual general shareholders' meeting.

Kyung articulated Samsung's ambition for its memory chip business to attain a greater profit share relative to its market share in the ongoing year. Samsung currently commands a 45.5 per cent market share in DRAM chips, which are essential components in various tech devices, according to data from TrendForce's fourth-quarter report.

To realise this objective, Samsung aims to secure a competitive edge in high-end memory chips, driven by the surging demand for artificial intelligence applications. This includes plans to mass-produce a 12-stack version of high-bandwidth memory (HBM) chips called HBM3E. Looking ahead to the anticipated release of HBM4 chips in 2025, Samsung intends to leverage its integrated capabilities in memory chips, chip contract manufacturing, and chip design to deliver more tailored solutions to meet customer demands.

Responding to a shareholder inquiry regarding Samsung's recent challenges in the HBM market compared to competitor SK Hynix, Kyung assured that measures are in place to prevent such setbacks in the future.

Following Nvidia (NVDA.O) CEO Jensen Huang's announcement affirming the qualification of Samsung's HBM chips for use in AI semiconductors, Samsung Electronics shares surged by 6.04 per cent on Wednesday. This marks the highest one-day increase since early September.

Additionally, Kyung highlighted Samsung's expectation of yielding tangible results shortly from other memory products under development for AI applications, including compute express link (CXL) and processing-in-memory (PIM) products.

(With Reuters inputs)