Memory chip shortage expected to persist until 2028 as AI demand outpaces manufacturing

Memory chip shortages show no signs of easing, with executives from major manufacturers forecasting constrained supplies through at least 2028. The supply crunch is being driven by surging demand for specialized chips used in AI systems and data centers, which are consuming the bulk of available production capacity.

Supply-demand gap widening#

Micron's CEO Sanjay Mehrotra indicated that his company's ability to fulfill orders will remain strained for the next two to three years. The executive noted that "overall, supply-demand environment is only getting tighter," reflecting competition between different memory types for limited manufacturing resources.

The shortage carries consequences beyond the server and AI sectors. Although Micron itself no longer manufactures consumer RAM, the shift in production toward high-bandwidth memory (HBM)—chips designed for artificial intelligence applications—and enterprise DRAM is leaving consumer devices starved for components. Manufacturers are dedicating their most advanced fabrication facilities to serve the lucrative AI market, effectively squeezing out traditional consumer-grade memory production.

Manufacturing bottlenecks ahead#

Micron is investing in new manufacturing facilities expected to come online in 2028, but the company does not expect these additions to immediately relieve the shortage. Mehrotra explained that ramping up production in new fabrication plants is inherently gradual, with complex technical challenges extending timelines even after first production runs begin.

The transition to newer chip architectures—moving from generation 3E to 4 and 4E variants of high-bandwidth memory—also complicates capacity gains. These advanced nodes yield fewer chips per silicon wafer than previous generations, creating a headwind for supply growth. With 75 percent of Micron's 2027 output already committed to existing customers, the company's near-term flexibility is severely limited.

AI chips crowding out other applications#

Demand for artificial intelligence-focused memory is outpacing traditional data center DRAM orders, further skewing production toward the highest-margin segments. This divergence deepens the squeeze on consumer and general-purpose enterprise applications, which must wait longer for available capacity.