DDR5 Prices Surge 5-10x, Obsolete DDR4 Memory Repurposed for AI Servers

Summary

Over the past year, memory prices have skyrocketed, with DDR5 modules now costing 5-10 times more than they did last year in some cases, leaving even AI giants reeling from the cost. As AI servers require massive amounts of high-spec DDR5 or LPDDR5X memory, supply constraints and soaring prices have made it unsustainable. In response, a trend of "DDR4 repurposing" has emerged: billions of DDR4 modules previously retired by server makers are being given a second life in AI platforms. Notably, Meta has developed Vistara technology, which uses CXL controller chips to connect DDR4 memory to platforms that natively require DDR5, allowing millions of decommissioned DDR4 sticks to avoid being scrapped and instead contribute to AI workloads. This development matters for gamers and the industry alike: soaring memory costs directly increase PC build prices, impacting hardware affordability. However, reusing DDR4 helps alleviate the supply-demand crunch for premium memory, potentially curbing runaway price inflation. It also showcases how legacy hardware can thrive in the AI era, offering new pathways for memory upgrades in servers and consumer PCs. For gamers, this means higher budgets for memory in the near term, but the industry's more efficient use of existing resources could stabilize prices in the long run and accelerate adoption of technologies like CXL.