Memory chip prices have climbed sharply as demand from large AI and cloud infrastructure projects has outpaced production, raising concerns of an emerging 'chipflation' that could ripple through hardware makers, cloud operators and consumer markets, Morgan Stanley said.
In a detailed 66-page report, the brokerage noted memory prices have surged about six-fold in the last year as producers concentrated on higher-margin parts for data centers rather than chips used in everyday devices. "What began as an AI infrastructure bottleneck is now spreading into hardware margins, device affordability, cloud costs, inflation and policy," the analysts wrote, adding that the crunch has "become a macroeconomic concern."
The note highlights how large technology companies' spending on AI infrastructure has reshaped memory demand. Producers have prioritized data-center grade products to capture better margins, leaving traditional device makers with a smaller, tighter supply of commodity memory. That reduced availability is forcing device manufacturers to decide whether to pass rising component costs on to customers or accept thinner profitability.
While some chipmakers are moving to expand capacity, Morgan Stanley cautioned that building new semiconductor manufacturing plants is costly and complex, and likely to take years before materially expanding supply. Unlike earlier cycles in the semiconductor industry, the brokerage suggested this episode could mark a longer-term supply-demand realignment because big cloud and AI firms are tying up capacity through long-term agreements and other commitments.
That hoarding of capacity, Morgan Stanley said, constrains the pool of chips available to more traditional buyers and increases volatility in supply. The brokerage pointed to knock-on effects showing up in producer-price measures, corporate margins, capital expenditure plans, cloud operating costs and delays in deploying new consumer technologies.
Several consumer electronics companies have already reacted. The report notes that makers from PlayStation producer Sony Group to PC manufacturer Lenovo have raised device prices, while some major cloud providers have signalled they will incur billions of dollars in extra outlays due to higher memory costs. Microsoft, for example, said in April that roughly $25 billion of its $190 billion in planned spending this year would reflect higher chip prices.
Research firm IDC is cited in the report as forecasting steep contractions in both the PC and smartphone markets in 2026 if rising prices deter buyers, particularly in lower-priced segments. Within the industry, Morgan Stanley described a clear divergence in fortunes: "Memory producers benefit from stronger pricing, margins and visibility. Downstream hardware companies must absorb costs, pass them through, redesign products or risk demand destruction."
The brokerage singled out producers of dynamic random-access memory controlled by a small group of suppliers. Samsung Electronics, SK Hynix and Micron together account for almost 90% of global memory output, the note said, and the shares of these companies have more than tripled this year. Geopolitical strains between the United States and China, together with export restrictions, are also fragmenting supply chains and tightening availability.
Morgan Stanley added that government subsidies are unlikely to provide fast relief because financial incentives do not overcome the long lead times required to bring new manufacturing capacity online. As a result, the current supply picture could remain constrained for several years even as firms attempt to add capacity.
Implications and observed effects
- Producer prices and corporate margins are already reflecting higher memory costs.
- Cloud service providers face increased operating expenses tied to elevated memory prices.
- Device affordability is under pressure, with some consumer electronics makers raising prices in response.
The brokerage's account frames the memory shortage as more than a sectoral supply disruption; it is presented as a cross-cutting issue that could influence pricing, investment and the pace of technological adoption across multiple parts of the economy.