Industry sources reported this week that Samsung Electronics confirmed a DRAM price increase of over 100% for its customers, marking the latest and most aggressive escalation yet in what has become a full-blown global memory crisis. According to Etnews, Samsung finalized supply price negotiations with several overseas customers, some of whom had already completed payment. The additional price hike, the company indicated, reflects price fluctuations between January and February.
This comes right after Samsung and SK Hynix had already pushed for server DRAM price increases of up to 70% for the first quarter of 2026, compared to Q4 2025 levels, according to Korea Economic Daily. Combined with the 50% increases that took place throughout 2025, prices have nearly doubled in the span of a year, and analysts say the worst is not over.
DDR5 contract pricing has now crossed the 100% increase threshold, reaching around $19.50 per unit compared to roughly $7 earlier in 2025. That G.SKILL DDR5 32GB kit that sat comfortably at $87 in early 2025 was already hitting $399 on Amazon by mid-December. And that was before Samsung’s latest round of confirmations.
The tech industry has started calling this the RAMpocalypse, and for good reason.
AI datacenters are consuming the world’s memory supply
The root cause of all this is no mystery. AI infrastructure demands an almost incomprehensible amount of memory. Nvidia’s latest Blackwell architecture ships with 192 gigabytes of RAM per chip, six times what a powerful gaming PC needs. One single NVL72 rack-scale system uses enough memory for a thousand high-end smartphones or several hundred beefy desktop PCs.
Now multiply that by the hundreds of thousands of units that Microsoft, Amazon, Meta, and Google are ordering every year. According to TrendForce, AI datacenters are expected to consume 70% of all high-end DRAM production in 2026. And since Samsung, SK Hynix, and Micron together produce about 90% of the world’s supply, when those three companies shift their factories toward high-bandwidth memory for AI accelerators, there is simply not enough left for everything else.
The numbers behind the AI spending are staggering. Meta, Microsoft, Amazon, and Alphabet collectively spent around $360 billion on datacenters in 2025, up from $217 billion in 2024. In 2026, that figure is estimated to reach $650 billion. These companies are paying top dollar to secure supply before anyone else, and they have the budgets to do it.
SK Hynix announced during its October earnings call that its HBM, DRAM, and NAND capacity is essentially sold out for all of 2026. Micron exited the consumer memory market entirely to focus on enterprise and AI customers. And Samsung sent a blunt message to its downstream customers in mid-December: “no stock.”
Procurement teams from Microsoft, Amazon, Google, Apple, and Dell have reportedly been flying to South Korea, packing business hotels in Pangyo, near Samsung and SK Hynix headquarters, trying to secure whatever allocation they can find.
The damage is already reaching regular consumers
This might look like a problem that lives inside server rooms and corporate boardrooms, but the impact is already hitting everyday buyers.
CyberPowerPC announced price increases on all its prebuilt systems in December, citing a 500% surge in global memory costs. Framework raised its DDR5 memory upgrade prices by 50% and warned that further increases were highly likely. Memory now accounts for roughly 20% of the hardware cost of a laptop, up from between 10% and 18% in the first half of 2025. HP reported that memory represents about 35% of what its PCs cost to build, double what it was just a year ago.
Analyst firm Gartner estimates that DRAM and SSD prices will surge 130% by the end of 2026, which would translate into a 17% increase in PC prices and a 10.4% drop in shipped units globally. Gartner also projects the sub-$500 entry-level PC segment could disappear entirely by 2028. Even Nintendo has publicly flagged memory cost pressure as a challenge for its Switch 2 projections.
The situation is not sparing smartphones either. Apple CFO Kevan Parekh tried to downplay it in October, calling it a “slight tailwind,” but Tim Cook has since acknowledged that the shortage will compress iPhone margins. The Phison CEO reportedly warned that many consumer electronics manufacturers could go bankrupt or exit product lines entirely by the end of 2026 if conditions do not improve.
On the contracts side, memory manufacturers have now abandoned annual supply agreements in favor of quarterly deals, giving themselves room to raise prices every three months. Samsung and SK Hynix are reportedly anticipating stepwise increases each quarter through at least 2027.
As for relief, most analysts agree it is still years away. TrendForce’s senior research VP Avril Wu noted that new production capacity from the major manufacturers will not make a noticeable difference in global supply until 2028. The factories are being built, but semiconductor fabs take years to come online, and AI demand is growing faster than anyone can keep up with.
What do you think, is this memory crisis getting completely out of control, or is this just the cost of the AI revolution? Tell us in the comments!