TrendForce reported that DRAM industry revenue rose 59.5% quarter over quarter in the second quarter of 2026, reaching close to $154.73 billion, attributing the jump to sharp increases in traditional DRAM contract prices as supply expansion continued to lag demand growth.
The increase reflects a broader pattern research firms have been documenting through the year: AI data center buildouts are absorbing a disproportionate share of global memory output, pushing prices up across both DRAM and NAND.
Why it matters
Memory pricing affects far more than data centers. JPMorgan's research published in August 2026 described the shortage as a self-reinforcing dynamic with knock-on effects reaching inflation and even cybersecurity risk, since memory cost increases ripple into consumer electronics, automotive systems and enterprise IT budgets.
The Motley Fool, citing industry estimates, reported that AI data centers are expected to consume roughly 70% of all memory chip production in 2026, effectively rationing what's left for phones, laptops, cars and other devices that also depend on DRAM and NAND.
How it works
Memory manufacturers allocate wafer capacity between conventional DRAM, NAND flash, and high-bandwidth memory used in AI accelerators. Because HBM commands far higher margins, producers including SK hynix, Samsung and Micron have shifted a growing share of their advanced capacity toward it, which reduces the wafers available for standard DRAM even as overall demand for computing memory keeps rising.
That reallocation is a major reason conventional DRAM prices have also spiked, not just HBM: TechInsights' June 2026 analysis noted the imbalance is being driven by AI rather than the usual PC or smartphone cycles that caused past memory shortages, making the current cycle harder to predict using historical patterns.
The competing read
Deloitte's July 2026 analysis was notably more pessimistic than some market commentary, arguing the shortage is worse than expected and may not ease until 2029, given how much new fab capacity takes years to bring online even after producers commit capital.
Other analysts frame this as a temporary supercycle likely to correct once new HBM and DRAM capacity from Samsung, SK hynix and Micron comes online in 2027 and 2028, closer to the historical pattern in which memory shortages eventually trigger oversupply once every producer expands at once.
What happens next
Contract price negotiations for the remainder of 2026 and into 2027 will be a key indicator of whether the shortage is easing or deepening, particularly as PC and smartphone makers, who compete with data centers for the same wafer supply, begin passing higher component costs to consumers.
Capacity additions announced by Samsung, SK hynix and Micron this year will start coming online in stages through 2027 and 2028, and how much of that new capacity goes to HBM versus conventional DRAM and NAND will shape how quickly, or whether, prices for everyday electronics come back down.