Samsung Posts Massive Q2 Profit on Memory Demand

Reuters reported that Samsung Electronics is likely to estimate second-quarter operating profit of 86 trillion won, or about $56.35 billion, based on an LSEG SmartEstimate of 30 analysts. Reuters said that would be about an 18-fold increase from 4.7 trillion won a year earlier and a third consecutive quarter of record operating profit for Samsung.

The report attributes the surge to prolonged memory shortages as demand from AI inference infrastructure continues to outpace supply growth. Reuters also reported that analysts expect the memory market to remain undersupplied at least through next year.

Reuters, citing Citi Research, reported that average selling prices for DRAM and NAND rose 44% and 53% quarter-on-quarter in the second quarter. The same report says demand is broadening beyond HBM into conventional DRAM and NAND as agentic AI and other inference workloads require more server memory and storage. Samsung's own first-quarter materials also framed record memory performance around high-value AI demand and limited supply availability.

The engineering implication is cost discipline. Memory-efficient inference, quantization, sharding, caching strategy, storage tiering, and workload placement become more valuable when the input market is tight. Buyers should also treat memory-heavy cloud-instance availability as a planning risk, not just a procurement detail.

• •Samsung's formal quarterly results and guidance later this month.
• •HBM shipment commentary from Samsung, SK Hynix, and Micron.
• •Contract and spot pricing for DRAM and NAND in the July-September quarter.
• •Cloud pricing or capacity changes for memory-heavy AI workloads.