Samsung expects the global memory-chip shortage to intensify through 2027 and remain tight into at least 2028 as AI infrastructure consumes a growing share of available production.
The company manufactures roughly one-third of the world's memory chips, giving its forecast weight across the technology supply chain. Samsung said frontier AI labs are sharing medium and long-term demand estimates directly as they compete to secure future capacity.
AI buyers are reshaping the market
Large customers are increasingly willing to sign long contracts for high-bandwidth memory and other components used in accelerators and data-center systems. That visibility lets Samsung justify new equipment and production expansion without taking the same risk that demand will disappear immediately after capacity comes online.
The memory business has traditionally moved through sharp boom-and-bust cycles. Manufacturers add capacity when prices rise, only to face oversupply later. Multi-year AI commitments could soften that pattern, although building fabrication capacity remains slow and expensive.
Higher prices are already producing mixed results inside Samsung. Its semiconductor division reached record sales during the second quarter, according to the company. At the same time, more expensive components reduced profitability in its smartphone and television businesses.
The pressure will not stay limited to AI companies. Memory is required in phones, laptops, game consoles, networking equipment, cars, and industrial systems. If manufacturers prioritize the highest-paying data-center customers, consumer-device makers may face higher bills or redesign products around more limited memory configurations.
A shortage lasting into 2028 would make memory one of the clearest links between the AI infrastructure boom and everyday electronics prices. Compute demand is no longer only changing cloud budgets. It is competing directly with the chips used across the rest of the technology industry.
