Stock Markets June 13, 2026 09:04 AM

Memory Price Surge Sparks Structural Supply Crunch, Morgan Stanley Says

AI-driven demand and capacity choices by DRAM suppliers are reshaping wafer allocation and leaving consumer markets increasingly undersupplied

By Sofia Navarro
Share
Twitter Reddit Facebook LinkedIn

Memory chip prices have spiked more than sixfold in the past year, a shift Morgan Stanley analysts describe as structural. Forecasts cited in the analysts' note project the memory market to expand from roughly $220 billion in 2025 to about $890 billion in 2026, driven primarily by price increases tied to surging demand for AI-focused memory such as high-bandwidth memory (HBM). Capacity constraints and leading-edge wafer allocation decisions by a concentrated set of suppliers are producing shortages across consumer and enterprise segments.

Memory Price Surge Sparks Structural Supply Crunch, Morgan Stanley Says
Summarize with
ChatGPT Perplexity Claude Grok Gemini

Key Points

  • Memory prices have climbed more than sixfold over the past year, reversing decades of steady price declines.
  • Forecasts cited in the note show the memory market growing from about $220 billion in 2025 to roughly $890 billion in 2026, a near $670 billion increase driven mainly by price rather than volume.
  • Concentration among three DRAM suppliers, capacity limits from EUV tools and wafer availability, and the rapid expansion of HBM for AI are creating supply imbalances that disproportionately affect consumer markets such as PCs and smartphones.

Memory chip pricing has reversed decades of steady decline, increasing more than sixfold over the past year and creating what Morgan Stanley equity analysts describe as a structural supply crisis rather than a cyclical blip.

TrendForce forecasts cited in the Morgan Stanley note put the memory market at roughly $220 billion in 2025 and project it to swell to about $890 billion in 2026. That roughly $670 billion year-over-year increase, the analysts observe, exceeds the standalone addressable market for whole industries such as smartphones, personal computers and servers.

For much of the modern computing era, the cost of a gigabyte of DRAM declined predictably. According to Morgan Stanley analysts Shawn Kim and Joseph Moore, that long-term period of roughly tenfold price declines every five years from 1957 to 2020 has ended.

The note attributes the reversal to demand rather than supply-side oversupply. Artificial intelligence workloads have produced a sudden, price-inelastic jump in demand for memory. The report highlights that supply is slow to react because adding capacity involves a multi-step process - installing new tools, qualifying production processes and ramping yields - that can take around two years.

Market concentration is an important factor. Three DRAM manufacturers control about 90% of DRAM supply and virtually all high-bandwidth memory, known as HBM. Those producers are prioritizing scarce leading-edge wafers toward higher-margin products such as HBM and server DRAM, reducing the share available for consumer markets.

Bit supply is expanding at roughly 30% annually but is constrained by limited extreme ultraviolet lithography tools and wafer capacity. At the same time, revenue in the memory market is rising by approximately fourfold. "Almost the entire spend surge is therefore price, not volume," the analysts said.

Global DRAM wafer capacity is expected to grow by about 30% by 2027, yet the industry focus on AI applications means consumer segments may remain undersupplied. Morgan Stanley’s projections suggest PC DRAM could be about 15% short of demand by 2027, equal to around 58 million units, while smartphone DRAM may face a 12% deficit, roughly 134 million units.

HBM, which is central to AI server performance, consumes substantially more wafer capacity per usable bit than conventional DRAM. The note states HBM uses three to four times the wafer capacity per usable bit and that its share of leading-edge memory wafer capacity could rise from about 6% in 2023 to 34% by 2028.

The analysts quantify the pricing pressure required to maintain gross margins if cost increases were passed through directly. Average selling prices would need to climb roughly 34% for smartphones, 67% for PCs, 83% for servers and 14% for storage products to offset higher memory costs alone. They add that a 67% rise in PC average selling prices would be the largest such increase on record, surpassing pandemic-era price moves by a factor of eight to nine.

But the broker notes that manufacturers, retailers and OEMs do not typically pass the full cost through to end consumers. "Nobody passes all of this through, which is why it surfaces instead as margin pressure, spec cuts and delayed launches," the note said.

Memory cost increases are not fully captured in headline consumer price indexes because memory is an intermediate input embedded in capital goods and services. The report points out electronic components have increased about 30% year-on-year in producer prices, according to the analysts.

Finally, the analysts describe the emergence of a two-tier market. Large cloud providers and hyperscalers are securing capacity through long-term agreements and prepayments, locking in supply, while other purchasers compete for a smaller and more volatile residual pool of memory.


Implications for markets: The note highlights how allocation decisions and capacity constraints tied to AI demand are altering the dynamics between enterprise and consumer segments, with consequences for product availability, corporate margins and the structure of supplier agreements.

Risks

  • Persistent undersupply in consumer DRAM markets - PC and smartphone DRAM are projected to face shortfalls (approximately 15% and 12% respectively by 2027), which could pressure device availability and margins in the consumer electronics sector.
  • Margin compression for original equipment manufacturers and vendors - because not all increased memory costs are passed to end customers, manufacturers may face margin pressure, leading to specification reductions or delayed product launches across technology hardware sectors.
  • Two-tier supply dynamics - hyperscalers securing capacity through long-term deals and prepayments could leave smaller customers competing for a limited residual pool, increasing volatility for downstream industries reliant on memory components.

More from Stock Markets

No Clear Dominant in Quantum Computing, Analysts Say; Pure-Play Firms Hold Modest Slices of a Large Future Market Jun 13, 2026 SpaceX IPO Forces Reconsideration of 'Magnificent Seven' Label as Market Roster Expands Jun 13, 2026 Two phones, a VPN and a state app: How Russians navigate tighter internet controls Jun 13, 2026 Indian regulator warns Tata Electronics plant after alleged wastewater seepage taints farm wells Jun 13, 2026 Bullion Boom Drives Some Gold Watches to the Crucible Jun 13, 2026