Samsung Electronics Q2 Profit Expected to Surge 18-Fold YoY to Record High
Claire Weston
Samsung Electronics' Q2 operating profit is forecast at ₩8.6 trillion, roughly 18 times last year's figure and a third consecutive record — driven by AI inference demand pushing memory chips into shortage and prices sharply higher.
Where does an 18× jump come from?
The core driver is AI inference infrastructure buildout. Servers need more memory and more storage. This means → not just HBM — high-bandwidth memory designed for AI chips — but ordinary DRAM and NAND are riding the same wave.
Citi Research data: Q2 average selling prices rose 44% quarter-on-quarter for DRAM and 53% for NAND. In plain terms = nearly every category of memory chip is repricing upward by roughly half.
Samsung's year-ago Q2 profit was just ₩4.7 trillion, when the memory industry was barely out of its downturn. This reflects a full cycle swing from trough to historic peak.
How long can the price rally last?
Analysts broadly expect supply shortages to persist at least through next year.
Nomura forecasts Q3 commodity DRAM prices up another 24% and NAND up 25%, supported by both consumer storage and data-center demand.
This means → memory makers hold strong pricing power in the near term. Samsung, SK Hynix, and Micron have gained 158%, 273%, and 242% respectively this year; all three now carry market caps above $1 trillion.
How much profit will employee bonuses absorb?
Samsung reached a labor deal in late May: 10.5% of the semiconductor division's operating profit goes to chip-unit staff as a special bonus.
Some analysts estimate cumulative bonus accruals could exceed ₩40 trillion. In plain terms = the higher the profit, the larger the bonus pool — and the timing of the accrual directly shapes the Q2 income statement.
Samsung will release detailed earnings later this month. Bonus accrual size and timing will be among the market's closest-watched variables.
If AI spending slows, who is hit first?
JPMorgan notes that investors broadly agree memory supply-demand is tight, yet many question the sustainability of AI storage's share of cloud capex — projected at 52% this year and above 70% next year.
This means → if major cloud providers cut AI spending, memory prices and demand would fall in tandem, hitting Samsung and SK Hynix hardest.
The two companies last week announced plans to invest ₩3,200 trillion (≈$2.07 trillion) in Korean chip capacity from 2026 to 2040. This reflects a bet that AI demand is a secular trend, not a short-lived bubble — if they are right, the capacity secures pole position; if wrong, it becomes overcapacity.
Content is for reference only, not financial advice.