PC memory costs surge 400% as Samsung, SK Hynix, Micron prioritize AI
Samsung, SK Hynix, and Micron are prioritizing HBM production for AI, leading to a 400% rise in PC memory costs. DRAM prices surged up to 98% in Q1 and 63% in Q2, while NAND flash prices rose up to 75%. HBM capacity is sold out through 2026, with new fabs expected to come online in the next two years.

*this image is generated using AI for illustrative purposes only.
A mainstream PC memory kit that cost about $75 a year ago now sells for as much as $460, driven by a strategic shift in manufacturing priorities rather than supply chain disruptions. Samsung Electronics Co., Ltd., SK hynix Inc., and Micron Technology, Inc. are allocating production capacity toward high-bandwidth memory (HBM) used in artificial intelligence (AI) accelerators and hyperscale data centers, reducing supply for consumer electronics. This reallocation has resulted in significant price increases for conventional DRAM used in laptops, desktops, and smartphones.
AI Drives Capacity Allocation
The three companies control the majority of the global DRAM market, giving them significant influence over production direction. HBM commands higher prices and consumes more manufacturing capacity per wafer compared to conventional DRAM. According to semiconductor commentator Shanaka Fernando, no coordinated action is required to tighten the market; manufacturers are simply following the higher margins offered by AI memory.
Price Surge Across Memory Segments
Data from TrendForce highlights the magnitude of the price increases. Conventional DRAM contract prices surged between 93% and 98% in the first quarter, followed by an additional rise of 58% to 63% in the second quarter. NAND flash prices also increased by 70% to 75% as suppliers prioritized AI-related products over mainstream memory.
| Memory Type | Price Increase Q1 | Price Increase Q2 |
|---|---|---|
| Conventional DRAM | 93% - 98% | 58% - 63% |
| NAND Flash | 70% - 75% | - |
Industry Impact and Future Outlook
The capacity constraints are affecting major technology companies. Tesla Inc. CEO Elon Musk warned of a "chip wall" due to AI infrastructure strain, while Dell Technologies Inc. CEO Michael Dell noted an explosion in memory requirements. HBM capacity is reportedly sold out through 2026, with much of 2027 production already committed. While new fabs are under construction in Idaho and New York, and Samsung and SK Hynix are expanding capacity, these additions are not expected to alleviate immediate shortages. The fundamental shift in capital allocation toward AI suggests consumer electronics will continue to compete for remaining capacity.
How will sustained high memory prices impact consumer demand for laptops and smartphones in the coming year?
Will the scarcity of conventional DRAM force PC manufacturers to develop alternative architectures or reduce standard memory specifications?
At what point might regulatory bodies scrutinize the major manufacturers' shift from consumer to AI-focused production?

































