Morgan Stanley warns: old storage prices are peaking, supply and demand may turn relaxed by 2028

The pricing premium in the legacy storage chip market is gradually coming to an end. Morgan Stanley warned in its latest report that the growing uncertainty around cloud capital expenditures, persistent weakness in consumer demand, and an anticipated supply-demand balance around 2028 mean that the profitability logic of the legacy storage industry has shifted from "upside elasticity" to "downside defense."

According to Trading Desk, Morgan Stanley's report noted that following the US midterm elections, it is expected that about 20% of new data center capacity slated for 2028 in the United States will be postponed to 2029, affecting construction totaling approximately 5 gigawatts, equivalent to 10% of the new global capacity for 2028.

This delay may create a "disruption" in the tech supply chain in 2028, compress the pricing power of legacy storage chips, and trigger inventory accumulation risks.

Meanwhile, after updating the supply and demand models for SLC NAND, MLC NAND, NOR, and DDR4, Morgan Stanley found that each segment is showing a trend toward equilibrium, especially pronounced by 2028.

The direct impact on the market has already emerged: Morgan Stanley downgraded several legacy storage-related stocks and sharply reduced their target prices, while also lowering 2027–2028 earnings expectations. Analysts are now using a 1.0x price-to-book ratio (P/B) as the downside anchor for valuation, rather than supporting the floor with peak earnings, marking a systemic shift in the market's pricing logic for the sector.

Pricing Advantage Becomes a "Double-Edged Sword" as Weakening Demand Exposes Vulnerability

Morgan Stanley's previous investment logic was founded on the differentiated pricing power of legacy storage chips—compared to mainstream storage, legacy storage is less constrained by long-term fixed-price agreements, allowing for quicker price hikes during periods of shortage. However, the firm now believes this flexibility is evolving into a two-way risk.

The report points out that visibility on cloud spending is declining, diminishing the certainty of incremental enterprise demand; meanwhile, consumer demand is sharply weakening, and end clients are clearly less able to absorb further price increases.

Morgan Stanley's research on PC semiconductors shows that there was a surge in concentrated "pull-in" orders from Q1 to Q3 2026, but inventory digestion deepened significantly in Q4, indicating that previous order volumes were higher than actual consumption demand, suggesting a risk of overdrawn demand.

The report specifically notes that the above analysis only concerns the legacy storage field, including DDR3/4, SLC/MLC NAND, and NOR flash, and does not involve the global mainstream DDR5 or TLC/QLC NAND markets.

Supply-Demand Model Update: Shortage Window Is Narrowing

Morgan Stanley's latest supply-demand models for SLC NAND, MLC NAND, NOR, and DDR4 all point to the same trend: the supply-demand gap is closing, and a basic balance or even oversupply is likely around 2028.

Specifically, forecasts show that the supply-demand gaps for SLC NAND, MLC NAND, and NOR in 2028 are about -3%, -5%, and -1% respectively, while DDR4 could see roughly a +1% supply surplus.

Analysts note that the current shortage may not disappear immediately, but the narrowing gap itself is enough to weaken incremental pricing power and trigger a re-rating of stock prices ahead of peak profitability.

Based on this, Morgan Stanley has lowered its ASP assumptions for several product lines for 2027 to 2028: ASP forecasts for NOR and DDR4 have been cut by approximately 23% to 26% in Q4 2028; ASP forecasts for some companies' SLC NAND have been cut even more significantly.

With shipment volume forecasts largely unchanged, revenue and gross margin estimates have been revised downward, and 2028 EPS forecasts have generally been cut by 10% to over 30%.

Data Center Capex Delays: 2028 Supply Chain "Cavitation" Risk

New data center legislation introduced after the US midterm elections is a key trigger for this shift in expectations.

According to calculations by Morgan Stanley's US team, legislative changes in states such as Texas and Ohio delayed about 5GW of data center construction from 2028 to 2029, which means new US data center capacity in 2028 is expected to decline by about 4% year-over-year, with a strong rebound of about 38% in 2029.

Globally, Morgan Stanley now forecasts that new data center capacity worldwide will increase by approximately 9% year-over-year in 2028, rising to around 20% in 2029.

This "low first, high later" cadence places phase pressure on the legacy storage sector—delayed demand combined with continuous supply buildup means a stage of oversupply is highly likely in 2028, when the risk of inventory cycles will rise significantly.

Notably, Morgan Stanley maintains its total capital expenditure forecast for global hyperscale cloud providers, expecting around $1.6 trillion in 2027 and $1.8 trillion in 2028, but clarifies that if buildout progress is further delayed, there is a risk of downward revision to the 2028 forecast.

Valuation Framework Shift: Using Book Value as the Downside Anchor

In the face of systematic downward earnings revisions, Morgan Stanley has adjusted its valuation methodology to use a 1.0x price-to-book ratio as the downside anchor, replacing the previous framework of using peak earnings as the valuation floor.

Analysts believe that for capital-intensive IDM and foundry enterprises, book value serves as psychological support at cyclical troughs—when earnings are temporarily pressured, investors often revert to the net asset basis for valuation.

At the individual company level, the magnitude of target price adjustments is quite significant, but the approach differs between IDM companies and fabless design firms.

The report makes it clear that the 1.0x P/B framework mainly applies to IDM/foundry firms, and not to fabless companies, as the latter lack the balance sheet valuation floor and therefore face higher downside volatility during the cycle.

Analysts identify price deceleration, order cuts, and inventory digestion as the main downside catalysts; a rebound in cloud order demand or a delay in supply expansion is seen as the primary upside risk in the current context.