DRAM

The lawsuit also raises concerns regarding how insulated the three big DRAM producers have become over time. Building even one DRAM fabrication plant costs billions of dollars and takes years, never mind the expertise required to keep one running. Although Chinese DRAM manufacturers like CXMT and YMTC are slowly picking up Steam, export and supply controls imposed by the US don't help the situation. The result? It's almost impossible for anyone to step in and compete with the big three manufacturers.

So what happened? AI happened. The way the world's memory market is set up relies on three major companies to supply everyone else with DRAM for consumer memory and storage. Samsung, SK hynix, and Micron realized early on in the AI boom that they could make a whole lot more money by moving away from consumer memory and into high-bandwidth memory (HBM) that's used in AI datacenters. Profit margins are enormous on the other side, and I really don't find it surprising that these massive corporations made the shift, as companies like Microsoft, Google, and Amazon have essentially offered blank che

Gartner expects PC lifetime for regular consumers to increase by 20% by the end of this year. The cheap PC upgrade cycle has all but come to an end. When new DRAM factories do come online, I expect them to serve AI customers first. Consumers will get the scraps, just like they are currently. Chinese DRAM makers might enter the international market and force prices to fall, but I doubt by enough to make a serious impact. 👉 Best Windows laptop in 2026 If you're waiting for the RAM crisis to be over before you buy your next device, I have some bad news. The most optimistic predictions put easing

As the global markets find themselves in the middle of a memory crisis, the obvious solution to a memory shortage is, well, more DRAM. The problem is that DRAM is becoming increasingly expensive. Memory alone is expected to account for roughly 30% of hyperscaler data center expenditure this year, which marks a fourfold increase since 2023. At those prices, simply throwing more memory at growing AI workloads is becoming an increasingly difficult proposition for cloud providers trying to keep infrastructure costs under control. That's the demand side of the issue. The supply side, on the other h