Anyone hoping for cheaper RAM in the near future might want to brace for disappointment. A new market analysis from Counterpoint Research suggests there is “no scenario” where memory prices meaningfully correct in the second half of 2027, signaling that the current surge in DRAM and NAND prices could persist for several years.
The warning comes amid what many analysts are calling a global memory crunch, driven largely by the explosive demand for AI infrastructure. Data centers building out large-scale AI systems require enormous amounts of memory, particularly high-bandwidth memory (HBM), which is consuming manufacturing capacity that would otherwise go toward consumer-grade chips used in PCs, phones, and SSDs.
Why are memory prices staying high for so long?
The core problem is straightforward: demand for memory is growing much faster than supply, largely due to the rapid expansion of AI infrastructure. AI data centers require enormous amounts of high-performance memory, and some estimates suggest they could consume around 70% of the world’s high-end DRAM by 2026.
At the same time, manufacturers such as Samsung, SK Hynix, and Micron are prioritizing higher-margin technologies like high-bandwidth memory (HBM), which are essential for AI accelerators. This shift is reducing the supply of conventional memory used in consumer electronics, pushing DRAM prices sharply higher and leading analysts to warn that volatility and price increases could persist as supply struggles to keep pace with demand.
What could this mean for PCs and smartphones?
For consumers, rising memory prices could ripple across the entire tech industry. Since DRAM and NAND are essential components in PCs, smartphones, GPUs, and storage devices, higher memory costs can quickly translate into more expensive hardware.
Analysts warn that sustained price increases could lead to double-digit price hikes for PCs and smartphones, forcing manufacturers to either raise prices or adjust specifications to protect margins. If current forecasts hold, the industry may be entering a longer AI-driven “memory supercycle,” meaning cheap RAM upgrades could remain out of reach for quite some time.