Samsung Posts First Profit Drop Since 2023 Amid HBM Chip Delays

Samsung Electronics reported a 56% drop in Q2 operating profit to 4.6 trillion won ($3.3 billion), its first decline in over a year. The company has struggled to sell AI memory to Nvidia due to certification delays, and is facing competition from SK Hynix and Micron. Samsung plans to begin mass production of HBM4 chips in H2.

Samsung Electronics reported a 56% drop in its Q2 operating profit, falling to 4.6 trillion won ($3.3 billion), its first decline in over a year. The company has been grappling with certification delays for its AI memory chips, particularly with Nvidia, and is facing stiff competition from rivals like SK Hynix and Micron [1].

The decline in profits is largely attributed to the slowdown in sales of advanced memory chips, including High-Bandwidth Memory (HBM) chips, which are crucial for AI data centers. Samsung has been experiencing difficulties in certifying its latest HBM3E 12-layer chips with Nvidia, leading to a delay in supply [2].

Ryu Young-ho, a senior analyst at NH Investment & Securities, noted that HBM revenue likely remained flat in the second quarter due to persistent China sales restrictions and the absence of significant shipments to Nvidia. He expects minimal shipments to Nvidia this year [1].

Samsung's reliance on the Chinese market for HBM sales has been further disrupted by U.S. export restrictions on advanced chips. The company's exposure to China accounts for roughly 20% of its HBM sales, and the potential revocation of technology authorizations for American chipmaking tools in China poses additional risks [2].

Despite these challenges, Samsung's smartphone sales are expected to remain robust, driven by pre-tariff stockpiling in the U.S. However, the company's overall performance has been underwhelming compared to its peers, with shares climbing about 19% this year but still trailing the benchmark KOSPI index by 27.3% [2].

Looking ahead, Samsung plans to begin mass production of HBM4 chips in the second half of the year, aiming to bolster its AI memory offerings. The company has also been working on securing orders for Nvidia GPUs in its advanced 2nm Gate All Around (GAA) process node, which promises improved power efficiency and performance [3].

The ongoing certification delays and regulatory risks have raised concerns about Samsung's competitiveness in the AI hardware market. As the company awaits the official release of its Q2 profit, investors remain cautious, with shares down nearly 2% as of Monday morning in Seoul [2].

References:
[1] https://economictimes.indiatimes.com/tech/artificial-intelligence/samsung-electronics-q2-profit-likely-to-drop-39-on-weak-ai-chip-sales/articleshow/122293202.cms
[2] https://techeconomy.ng/samsung-q2-profit-projected-to-drop/
[3] https://www.tweaktown.com/news/106178/samsung-foundry-could-make-nvidias-next-gen-gpus-on-its-in-house-2nm-process-node/index.html