RAMpocalypse
The “RAMpocalypse” refers to the global memory crisis that began in late 2025 triggered by the hype regarding Artificial Intelligence (AI), causing the price of RAM and NAND flash to skyrocket.
Many will remember that supply challenges during COVID as companies around the world were forced to investment in technology (e.g., laptops) to support home working. We then had the cryptocurrency boom, that limited the availability of Graphics Processing Units (GPUs), causing a surge pricing and extensive scalping.
The difference with the RAMpocalypse is that RAM and NAND flash is used in almost every modern electronic device, meaning the impact is extremely broad, covering consumer and business spending.
Most of the world’s RAM and NAND flash comes from three companies, specifically Samsung, SK Hynix and Micron.
These companies can still produce RAM and NAND flash, but they have shifted their production towards higher margin memory products required to support the scale of Artificial Intelligence (AI). This demand comes from the insatiable appetite of companies such as OpenAI, Anthropic, Microsoft, etc.
This is great for Samsung, SK Hynix and Micron. However, the shift in production means much lower supply of everyday RAM and NAND flash, resulting in significant price increases. In a perfect world, the consumer or (end buyer) could be protected from the price increase. Unfortunately, due to the scale of the increase (> 130%) this is not possible with the RAMpocalypse.
To highlight the impact, outlined below are four popular consumer products, including their approximate price increase.
- Apple MacBook Air = Increase of ~$200
- Apple MacBook Pro (M5 Max) = Increase of ~$500
- Sony PlayStation 5 = Increase of ~$90
- Xbox Series S = Increase of ~$100
In short, the more RAM or NAND flash used by a product, the higher the likely price increase.
Personally, I am pleased I ordered my Apple MacBook Pro M5 Max early, as the equipped 128GB of RAM has resulted in a $2000 increase.
The main concern is that it would appear this may be the new normal, as there is no evidence that the demand for Artificial Intelligence (AI) is set to slow.
In addition, the time and cost associated with scaling demand, for example, new manufactures entering the market is time consuming and costly (especially in am market with high price volatility).
As a result, we must acknowledge that buying technology, which has always been cost prohibitive, has become even more challenging.
Looking at a small positive. I do believe that constraints can force innovation. For example, in a world where RAM and NAND flash is restricted, software developers will be forced to think differently about their architecture, looking for ways to optimise performance and memory utilisation. It could also foce more to think about the right to repair, increasing hardware longevity.