- A report claims that memory prices could rise by 40% to 50% by the end of Q1 2026.
- Analysts indicate a "permanent reallocation" of RAM supply towards AI companies.
- Data centers are expected to consume over 70% of the total supply of high-end memory chips this year.
As the RAM crisis intensifies, a recent report highlights the potential long-term effects on consumers, reminiscent of the supply chain disruptions caused by the pandemic.
Tom's Hardware referenced an article from The Wall Street Journal (WSJ), emphasizing the severity of the RAM crisis and its expected escalation.
According to Counterpoint Research, memory prices are anticipated to increase by 40% to 50% by the end of March 2026, following a 50% rise in Q4 2025.
This situation is driven by the AI boom, as data centers require substantial amounts of RAM to support increasingly popular Large Language Models (LLMs) like ChatGPT, Copilot, and Gemini, among others.
The servers in these data centers demand vast quantities of RAM, and the powerful graphics cards essential for AI operations also require significant video RAM. Consequently, AI companies are investing heavily to sustain their growth.
The WSJ report indicates that the immense purchasing power of AI is sidelining other industries seeking memory, potentially affecting a wide range of technologies.
This includes not only the RAM in laptops, PCs, and smartphones, but also consumer electronics like TVs and cars—all facing potential price increases due to the disrupted memory supply caused by AI. This disruption is compounded by earlier decisions made by memory chip manufacturers to reduce production in response to a previous inventory surplus.
As Avril Wu, a senior research VP at TrendForce, states: "I have tracked the memory sector for almost 20 years, and this time really is different. It really is the craziest time ever."
Another analyst firm, IDC, predicts that memory and storage price increases will lead to a 9% decline in PC sales and a 5% dip in phone sales in 2026, as these costs are inevitably passed on to consumers.
Furthermore, IDC notes a "permanent reallocation" of supply favoring AI firms, adversely impacting all other sectors reliant on memory. It is projected that data centers will consume over 70% of the total supply of high-end memory chips produced in 2026.
Analysis: AI is Consuming Your RAM
Is there any hope for consumers facing the RAM-consuming AI surge?
Major memory chip manufacturers—Micron, Samsung, and SK Hynix—are ramping up the construction of new production facilities to increase chip output. However, these plans are long-term and won't alleviate the supply of RAM until 2028, aligning with predictions that the RAM crisis will persist through 2027.
In the meantime, innovative solutions may emerge, such as repurposing old memory chips. The WSJ highlights Caramon, a company that reclaims RAM from decommissioned servers, which has seen its sales nearly double since the memory crisis began.
Consumers might consider purchasing used RAM from auction sites or salvaging memory from old PCs for new builds as temporary measures.
There is also hope that the AI industry will find creative ways to reduce its dependence on large amounts of RAM, as seen in recent developments with DeepSeek.
For now, the reality of the AI RAM crisis poses significant challenges for consumers in the coming years.