memory chip shortage

basicwidespread

The market situation where AI data centers buy up so many memory chips that fewer and more expensive ones remain for phones, laptops, and other devices.

In plain terms

Every computer, phone, and server needs memory: this is the machine's 'work desk' where it holds the data currently in use. The most widespread type is DRAM. The memory chip shortage means that factories worldwide can produce less of it than buyers want to purchase.

Imagine it like a bakery where a large client reserves most of the ovens for months in advance for wedding cakes. The rolls would be baked in the same oven, so fewer are made and they become more expensive. Here, the cake is the special memory intended for AI servers, and the rolls are the traditional memory going into your laptop and phone.

This shortage differs from previous chip shortages because it is not caused by a pandemic or natural disaster. It is driven by sustained high demand from data centers built to serve artificial intelligence, so it will not disappear within a few months.

How it works

Three companies account for the vast majority of global DRAM production: Samsung, SK Hynix, and Micron. If they reallocate capacity elsewhere, the entire market feels the impact.
AI servers require special high-bandwidth memory (HBM) alongside their accelerator cards. Producing this consumes far more manufacturing space per unit of storage capacity compared to traditional memory.
Since AI-targeted memory offers higher margins and major cloud companies book it years in advance, manufacturers prioritize it. Consequently, less memory is produced for PCs, phones, and traditional servers.
Due to shrinking supply, contract and spot market prices have risen sharply, while device manufacturers respond with price hikes or models with smaller memory.
Building and commissioning a new chip factory takes several years and billions of dollars, so supply can only catch up with demand slowly.

An everyday example

You are looking at a new laptop and see that the same model costs much more than six months ago, or is sold with less memory for the same price. The main reason is usually not the store's profiteering: the manufacturer obtains memory chips at a higher price because AI data centers consume most of them.

Who is affected

everyonePhones, laptops, game consoles, and other smart devices may become more expensive or reach stores with less memory.

entrepreneur, executiveThe costs of corporate equipment procurement, server expansion and cloud services are rising, and delivery times may lengthen.

developerAI models and agents running on local machines require significant memory, making hardware upgrades more expensive and harder to plan. → Code

bank customer, investorShares of memory and device manufacturers fluctuate strongly based on shortages and price trends. → Money

A common misconception

MythThe memory chip shortage means factories have halted or are producing less than before.

In realityFactories are working at full capacity, but an increasing share of production goes into AI servers. The shortage stems from capacity reallocation and extraordinary demand.

Limits and risks

No one can reliably say how long it will last: forecasts depend on whether the momentum of AI investments persists.
The memory market is traditionally cyclical: shortages have previously been followed multiple times by overproduction and price drops, so current prices will not remain forever.
If the purchase can be postponed, it is worth considering, but there is no guarantee it will be cheaper in a few months.
The shortage did not stop at DRAM: it spread to NAND flash memory used in storage drives, so SSDs also became more expensive.

Similar, but not the same

global chip shortage (2020–2023) – The pandemic and supply chain disruptions caused it, affecting various chips and industries, most visibly the automotive sector. The current one is driven by AI demand and focuses on memory.
GPU shortage – GPU shortage refers to the scarcity of accelerator cards used for AI computing, while memory chip shortage refers to the scarcity of memory used alongside them and in all other devices.
my computer is low on memory – A setup or upgrade problem with a single device is one thing; the memory chip shortage, however, is a global supply problem.

Related concepts

Broader

chip shortage, semiconductor industry, AI infrastructure, supply chain

Narrower

DRAM, high-bandwidth memory (HBM), NAND flash

On the edge

data center, GPU, AI accelerator, computing capacity, AI investment wave

A brief history

The first commercial success of DRAM, Intel's 1103 chip, appeared in 1970, and the memory market has since been characterized by alternating shortages and overproduction. The current shortage driven by AI demand became widespread in the second half of 2025, when memory prices began to rise sharply.

Where things stand

By the end of 2026, scarcity is part of daily life: memory prices are high, and a large portion of manufacturers' capacity is tied up by AI data centers. New factories are being built, but most analysts expect meaningful relief only in the coming years.

Status: 2026 November

In the news

Appears in 1 Agent Age stories; the curve shows weekly occurrence in the news magazine.

AI data centres drive memory chip shortage as Micron turns to long-term contractsOct 19, 2026 · Ipar

Also known as

memory shortage · RAM shortage · DRAM shortage · memory crisis · memory chip shortage · DRAM shortage · RAM shortage · memory supercycle

Sources

This entry was machine-edited and has undergone machine expert review; the “In the news” section updates automatically from Agent Age stories.