If you’re planning to build a PC or browsing pre-built programs, you may have encountered the same issue: abnormally high RAM values. In some cases, costs have increased by as much as 500%, increasing system prices and making construction difficult to justify.
This is not a temporary change. It is the result of many shifts happening at once, most of which correspond to how quickly memory demand has changed.
The demand for AI is a major driver behind the rise in RAM prices
The rapid growth of AI, especially large language models, has greatly increased the demand for memory. These systems depend on RAM for both training and guidance. Training benefits from faster data access compared to traditional storage, while visualization requires memory to hold ongoing conversations and context in real time.
Each user interaction with the AI model requires memory allocation, even if it is temporary. When that demand reaches millions of users, the amount of memory required increases significantly.
Larger AI models increase memory requirements even more
Demand isn’t just growing because more people are using AI. It also increases because the models themselves require more memory.
Context windows, which determine how much information a model can process at once, have grown significantly. Where high-end systems once handled around 200,000 tokens, newer models are now pushing over one million. That change directly increases the amount of memory required for each interaction.
Consumer RAM competes with AI infrastructure
Although consumer PCs do not use the same type of memory as AI data centers, they rely on the same production system.
AI systems use high-bandwidth memory (HBM), which is fast and designed for large workloads. However, HBM still relies on the same silicon wafers, manufacturing facilities, and supply chains used to produce standard DRAM for consumer RAM modules.
Because AI companies are willing to pay more, manufacturers prioritize that demand, which reduces the availability of consumer products.
Supply constraints tighten the market
On the supply side, there are only three major players: Micron, SK Hynix, and Samsung. That puts a limit on how fast the product can grow.
Micron has already shifted its focus from consumer RAM by shutting down its Core brand, reallocating resources towards AI demand. At the same time, SK Hynix has reportedly sold its 2026 production capacity, further limiting supply.
These constraints create a clear imbalance of supply demand, which further increases prices.
Why RAM prices may drop soon
Although manufacturers plan to increase production capacity, those changes take time. New artificial plants take years to build and bring online.
Current expectations suggest that a meaningful increase in supply may not reach consumers until 2027 or 2028. Meanwhile, some projections indicate that prices may rise significantly, possibly by another 100% from current levels.
This makes the short-term outlook uncertain for consumers hoping for immediate relief.
There are early signs that demand may be easing
Some improvements could ease the pressure on the market, but it’s early days.
Google introduced a method called TurboQuant AI compression, which aims to reduce the memory requirements of AI models up to six times. If similar methods are widely adopted, they can reduce the overall demand for memory in data centers.
The market’s reaction to the announcement suggests that manufacturers are paying attention, although the long-term impact remains unclear.
What PC builders should be doing right now
For consumers, the decision depends on priorities and timing.
If cost is the main concern, building older platforms that currently use DDR4 memory is the most viable option. DDR4 remains very affordable and still widely available, especially in the used market.
For those planning a high-end build with the latest hardware, waiting may not lead to better prices in the near term. In fact, prices can rise significantly before settling, making it difficult to predict the right time to buy.
Will RAM prices drop again?
Even if the supply improves, RAM prices may not return to their previous lows.
Markets tend to reset after a sustained increase in demand, especially when prices normalize for both producers and consumers. As AI continues to drive greater demand and companies prioritize high-net-worth enterprise customers, the base cost of memory is likely to remain high compared to previous years.
Prices may ease from current highs as new production capacity comes online, but a full return to pre-surge levels seems unlikely in the near term.
