News report
DRAM and NAND Contract Prices Are Still Rising in Q4 2026
TrendForce projects conventional DRAM contract prices up 10–15% and NAND Flash up 15–20% QoQ in Q4 2026 as AI demand keeps supply tight.
On this page
TrendForce expects another quarter of memory price increases
TrendForce projects conventional DRAM contract prices to rise 10–15% quarter over quarter in Q4 2026, while NAND Flash contract prices are expected to increase 15–20%. The forecast was published September 30 and reflects contract-market conditions rather than a guarantee about any specific retail RAM kit or SSD.
The research firm says DRAM suppliers continue to prioritize advanced-process capacity for high-performance server products, keeping the broader market undersupplied. For NAND, AI infrastructure demand is strengthening higher-value enterprise storage even while consumer demand remains comparatively weak.
PC DRAM remains exposed to the server-capacity shift
For PC DRAM, TrendForce says notebook inventories increasingly reflect higher component costs and PC OEMs are still procuring aggressively because they expect supply to remain tight into 2027. The firm also warns that PC DRAM supply could decline next year as more manufacturing capacity shifts toward server applications.
That does not mean desktop DDR5 prices must move by the same 10–15% contract-market forecast. Module inventories, retail competition, regional pricing, product density and timing can all change what consumers actually pay. The useful signal is that upstream conventional DRAM pricing pressure remains positive rather than easing in Q4.
Client SSDs and enterprise SSDs are moving differently
TrendForce describes a split NAND market. PC makers built finished-goods and channel inventory earlier in the year and are reducing SSD capacities in some mainstream systems to control bill-of-material costs. That limits incremental client-SSD procurement and gives NAND suppliers less room to push client pricing as aggressively.
Enterprise SSD demand is much stronger. TrendForce projects enterprise SSD bit demand to grow by more than 80% year over year in 2026 as cloud providers expand AI inference infrastructure, real-time retrieval, caching and vector-database workloads. Most incremental enterprise supply is already committed, and TrendForce expects enterprise SSD price growth to accelerate in Q4.
AI demand is increasingly splitting the memory market by workload
The notable Q4 pattern is not simply that every memory product is scarce to the same degree. Server DRAM remains undersupplied, graphics-memory output is constrained as suppliers reduce GDDR6 and advanced capacity goes elsewhere, while consumer-facing demand is weaker in notebooks, smartphones and client SSDs.
For PC buyers, that makes actual street pricing more useful than trying to convert one upstream percentage into a purchase deadline. For the industry, however, the forecast reinforces a broader 2026 trend: AI infrastructure is influencing how suppliers allocate DRAM and NAND capacity, and those allocation decisions can indirectly affect consumer components even when end-market PC demand is soft.
Sources
Primary and technical sources
These sources support the reporting and analysis above. Current stories are updated when later evidence materially changes the facts.
Related
Continue from here
Useful next steps selected from the same technical reference and publication system.
Technical guide
DRAM vs DRAM-Less NVMe SSDs: HMB and Tradeoffs Explained
Understand SSD controller DRAM, DRAM-less NVMe designs, Host Memory Buffer (HMB), cache boundaries, and why exact drive design and workload matter more than one label.
Tool
DDR Memory Latency Calculator
Convert DDR data rate and CAS latency cycles into CAS timing in nanoseconds.
Technical guide
Windows Page File Explained: Virtual Memory and Commit Limit
Understand what the Windows page file does, how it extends the system commit limit, how paging differs from RAM use, and why crash dumps can depend on it.
Tool
DDR Memory Bandwidth Calculator
Calculate theoretical peak DDR memory bandwidth from transfer rate, bus width per channel, and active channel count.