'Memflation' explained: why AI demand is pushing memory-chip prices sharply higher

What 'memflation' means: Industry analysts have started using 'memflation' to describe the sharp, sustained rise in memory-chip prices triggered by AI data-centre buildouts overwhelming supply. It is not a CPI-style inflation measure — it is a sector-specific price index for DRAM (system memory) and NAND flash (storage), and it sits inside the broader $1.3 trillion semiconductor revenue forecast for 2026.

The numbers: Estimates put DRAM price increases at around 125 per cent and NAND flash at around 234 per cent over the current surge cycle, with no significant relief expected until late 2027. Within that, some segments saw quarter-on-quarter jumps of 80 to 90 per cent between the fourth quarter of 2025 and the first quarter of 2026. A leading global smartphone maker has reportedly budgeted for a roughly 25 per cent increase in DRAM cost per phone in its 2026 line — a hit that, if passed on, can add about 25 per cent to the bill of materials for a mid-range device.

Why is it happening: AI servers need large amounts of high-bandwidth memory (HBM), specialised stacks like HBM3E, and high-capacity DDR5. These are far more profitable for memory makers than commodity DRAM and NAND used in smartphones, PCs and consumer electronics. Manufacturers have therefore reallocated wafer capacity towards AI memory, tightening supply for everything else. Fab capacity cannot be added overnight — new lines take 18 to 24 months — so the squeeze persists.

India angle: India is in the early stage of its semiconductor manufacturing journey under the India Semiconductor Mission, with assembly and test plants by Tata, Micron and CG Power coming online and a Tata-PSMC fab under construction in Dholera. None of these immediately make the kind of HBM that AI servers need, so India remains a net importer of memory chips. Memflation therefore feeds into higher input costs for Indian smartphone assembly, PC manufacturing and electronics exports, with second-round effects on retail prices and the trade deficit.

Exam angle: Note the term (memflation), the headline rise (DRAM ~125 per cent, NAND ~234 per cent), the cause (AI data-centre demand and HBM/DDR5 reallocation), the global revenue forecast ($1.3 trillion in 2026), and link it to the India Semiconductor Mission, the Tata-Powerchip Dholera fab, and the broader Make in India electronics push.