The core issue is not that AI-related memory demand has disappeared. According to JPMorgan's July 14 report, the selloff reflects a shift from enthusiasm about AI infrastructure expansion to a stricter test of profits, cloud capital expenditure, HBM pricing, and contract quality. The most important variable is whether hyperscale cloud providers can keep raising capex expectations enough to support the earnings already priced into memory stocks.
| Primary source | Wallstreetcn |
|---|---|
| Reported at | 2026-07-14T13:32:28.000Z |
| Topic | 股票 |
| Evidence limit | Reported facts are separated from interpretation; current prices and platform terms require independent verification. |
Evaluate BINANCE for your use case
Check regional eligibility, current fees and product availability on the official destination.
Review BINANCEWhat Changed In The Memory Trade
The memory trade has moved into a verification phase. Earlier in the cycle, investors were willing to lift assumptions for AI data center spending and memory total addressable market. After a sharp rally, they are now asking whether those assumptions are ahead of actual cloud provider budgets.
JPMorgan says about 70% of current market sentiment is centered on one question: can hyperscale cloud service providers keep revising future capital expenditure sharply higher? If that momentum slows, memory shares may remain under pressure in the short term.
This matters for crypto-market readers because AI infrastructure, semiconductors, and risk appetite often sit in the same macro conversation as high-beta technology and digital-asset markets. The article does not establish a direct price link to Binance-listed crypto assets, but it does describe a broader risk-pricing debate.
The Three Pressure Points
JPMorgan identifies three main reasons for the correction. First, AI investment expectations may have run faster than actual spending plans. Many investors now expect global hyperscale cloud capex to be revised toward 1 trillion to 1.5 trillion dollars over the next 3 to 6 months. If earnings reports do not validate that optimism, sentiment could weaken further.
Second, DRAM price momentum is slowing. After a run of price increases, JPMorgan says year-over-year and quarter-over-quarter DRAM price gains began to moderate after the second quarter of 2026. That cools expectations for rapid profit expansion.
Third, Samsung Electronics earnings expectations were cut before second-quarter results. That pre-emptive reduction hurt investor confidence and shifted the debate from how fast the industry can grow to how long current profit levels can last.
Why Long-Term Agreements Are Still Debated
Long-term agreements, or LTAs, were one of the most discussed topics in JPMorgan's roadshow. Investor attitudes have improved compared with several months ago. The focus has shifted from whether LTAs exist to whether suppliers can use them to lock in core AI customers.
The disagreement has not disappeared. More than half of surveyed investors still take a cautious view because the coverage ratio of Korean suppliers' LTAs is not transparent and contract quality differs across companies.
JPMorgan expects more than half of future contract volume to move into LTA frameworks. The bank also argues that LTAs do not necessarily cap future pricing upside, because some new orders can be repriced later and take-or-pay terms can improve order certainty.
HBM Pricing Is The Biggest Gap
HBM pricing is the largest expectation gap in the report. Many buy-side institutions expect 2027 HBM selling prices per gigabyte to roughly double year over year, and they use that assumption to support further earnings upgrades.
JPMorgan is more cautious. It estimates the current industry average HBM selling price at about 1.8 dollars per gigabyte, even slightly below some high-end server DRAM products. The bank expects 2027 HBM average selling prices to rise 25% to 30% year over year, which is far below the more optimistic buy-side view.
The bank's reasoning is that memory suppliers and cloud customers do not negotiate HBM in isolation. They consider DRAM, NAND, and HBM profitability together, which may limit how far HBM pricing can rise. Still, HBM is generally repriced annually, so stronger-than-expected AI demand could leave room for later increases.
DRAM Tightness And Enterprise SSD Demand
JPMorgan remains relatively positive on memory supply and demand. DRAM is described as the tightest product area, with supply able to meet only about 50% to 60% of order demand. NAND's comparable figure is about 70% to 80%.
Even with future DRAM wafer capacity expansion, JPMorgan expects tight supply-demand conditions may last into 2027 to 2028. That is one reason the bank does not frame the correction as a simple industry downturn.
Enterprise storage is also becoming more important. JPMorgan says consumer NAND demand has been revised down more than expected, but AI data centers are lifting enterprise SSD demand, including use cases such as KV Cache Offload. The supply chain expects 2027 enterprise SSD shipments to approach 500EB, nearly 50% year-over-year growth, with possible room for further upward revisions.
Practical Checks For Readers
The next useful check is cloud capex guidance. If large cloud providers do not confirm stronger spending plans in future earnings reports, the memory sector's AI premium may face more pressure.
The second check is HBM pricing realism. A large part of the upside case depends on whether HBM prices can rise close to aggressive buy-side expectations or only within JPMorgan's lower 25% to 30% year-over-year range for 2027.
The third check is whether LTAs improve earnings stability without limiting upside. Contract coverage, contract quality, and take-or-pay terms are more important than the headline existence of LTAs.
Risk Disclosure And Binance Context
This article is based only on the supplied JPMorgan-related event brief and Wall Street News source summary. It does not verify the original report independently, does not add external market data, and does not claim any ranking, indexation, trading outcome, or traffic result.
The analysis is not financial advice. Memory shares, semiconductor equities, and crypto assets can all be volatile, and the brief itself warns that markets carry risk. Readers should evaluate whether any view fits their own objectives, financial situation, and risk tolerance.
For readers who follow market news through Binance, the practical value is context: AI infrastructure expectations can influence broader technology sentiment, but this brief does not show a direct causal impact on any specific crypto asset. If you use Binance, the supplied referral context is code 7nfg8123 at BINANCE official destination.
Evaluate BINANCE for your use case
Check regional eligibility, current fees and product availability on the official destination.
Review BINANCEAffiliate link · Availability varies by region · No guaranteed outcomeQuestions readers ask
Why did Asian memory stocks fall so sharply?
According to the brief, the fall was mainly an expectations reset. Investors are questioning whether cloud capex, HBM pricing, DRAM price momentum, and earnings durability can support the optimism already reflected in share prices.
Does JPMorgan say AI memory demand has disappeared?
No. The brief says JPMorgan views the correction as a shift from AI infrastructure expansion expectations to profit verification, not as evidence that memory demand has vanished.
What is the biggest uncertainty in the report?
The biggest uncertainty is hyperscale cloud provider capital expenditure. JPMorgan says about 70% of market sentiment is focused on whether future CSP capex can keep being revised sharply higher.
Why is HBM pricing so important?
HBM pricing directly affects earnings upside. Many buy-side investors expect 2027 HBM prices per gigabyte to roughly double year over year, while JPMorgan expects a more moderate 25% to 30% increase.
Are long-term agreements good or bad for memory suppliers?
JPMorgan views LTAs more as a tool for profit stability than a cap on profit upside. The debate remains because investors lack full visibility into contract coverage and quality.
What should market readers watch next?
The most useful checks are future cloud capex guidance, HBM average selling price trends, DRAM price momentum, Samsung earnings expectations, and whether enterprise SSD demand keeps being revised higher.