South Korean memory manufacturer SK Hynix expects the global memory shortage to persist for much longer than initially predicted. In its quarterly report, published on October 28th, the company states that demand for memory is expected to exceed production capacity until 2030.
The primary reason? No surprise: significant investment in AI data centers, which rely on ample amounts of RAM. The widespread and ongoing expansion of data centers will continue to gobble up enormous amounts of memory, while demand is expected to rise further as AI is incorporated into other industries like self-driving vehicles and robots.
Why not produce more memory? Manufacturers are working on it, but it takes several years at minimum to build new semiconductor factories and spin up production to necessary levels. SK Hynix also points out that the actual manufacturing processes for increasingly advanced memory chips have gotten more technically challenging in recent years.
SK Hynix has already signed long-term supply agreements covering more than half of its planned production capacity for 2027 and 2028. And earlier this year, SK Hynix announced plans to invest $13 billion into a massive fabrication facility, but that won’t become operational until 2030. Meanwhile, Samsung—one of the two other major memory makers—says its own shortage situation will last until at least 2028.
The ongoing shortage will continue to have consequences for ordinary consumers around the world. As manufacturers prioritize profitable enterprise memory chips for AI servers, production and supply of consumer-grade RAM and storage will stagnate or even dwindle, causing prices to rise even higher than they’ve already gone in 2026. Any device that relies on RAM and storage—from phones to laptops to GPUs to smart home gadgets—will feel the squeeze that hasn’t yet fully squoze.
