Samsung Takes Steps To Revamp Chip Manufacturing To Align With Nvidia’s AI Requirements: Report (UPDATED)

Editor’s Note: This article has been updated to clarify Samsung’s strategic update in its chip technology, aiming to better compete for the business of AI chip industry leaders, including Nvidia.

Samsung Electronics Co (OTC:SSNLF) is adopting a new chip-making technique, previously advocated by its competitor SK Hynix, to enhance its position in the high-end chip market, crucial for powering artificial intelligence.

This strategic shift aims to position Samsung more competitively, potentially securing deals with AI chip leaders like Nvidia Corp (NASDAQ:NVDA), which it has missed out on in the past due to sticking with older technology that led to production challenges, Reuters reports.

The demand for high bandwidth memory (HBM) chips, crucial for generative AI, has surged, placing Samsung in a position to catch up with SK Hynix and Micron Technology Inc (NASDAQ:MU), who have been more active in collaborations with Nvidia

K Hynix has gained a considerable advantage by becoming the first supplier of HBM3 chips to Nvidia, capturing over 80% of the market share for these advanced memory products.

By moving to the mass reflow molded underfill (MR-MUF) method, Samsung addresses these issues, signaling its readiness to meet the rising demand for high bandwidth memory (HBM) chips driven by generative AI’s growth.

Recent developments indicate that Samsung has issued purchase orders for equipment capable of handling the MUF technique, signaling a strategic shift to enhance HBM production yields, according to Reuters.

Samsung’s production yields for its HBM3 chips are currently estimated at 10-20%, significantly lower than SK Hynix’s 60-70% yield rates for its HBM3 production.