Futures
Access hundreds of perpetual contracts
CFD
Gold
One platform for global traditional assets
Options
Hot
Trade European-style vanilla options
Unified Account
Maximize your capital efficiency
Demo Trading
Introduction to Futures Trading
Learn the basics of futures trading
Futures Events
Join events to earn rewards
Demo Trading
Use virtual funds to practice risk-free trading
CFD
U.S. stock CFD derivatives
US Stocks
Access real US stocks and ETFs
HK Stocks
Trade quality Hong Kong-listed stocks
Korean Stocks
SK Hynix
Real Korean stocks and top assets
Stock Futures
High leverage, 24/7 trading
Tokenized Stocks
Backed by real stock assets
IPO Access
Unlock full access to global stock IPOs
GUSD
Mint GUSD for Treasury RWA yields
Stocks Activities
Trade Popular Stocks and Unlock Generous Airdrops
Launch
CandyDrop
Collect candies to earn airdrops
Launchpool
Quick staking, earn potential new tokens
HODLer Airdrop
Hold GT and get massive airdrops for free
IPO Access
Unlock full access to global stock IPOs
Alpha Points
Trade on-chain assets and earn airdrops
Futures Points
Earn futures points and claim airdrop rewards
Promotions
AI
Gate AI
Your all-in-one conversational AI partner
Gate AI Bot
Use Gate AI directly in your social App
GateClaw
Gate Blue Lobster, ready to go
Gate for AI Agent
AI infrastructure, Gate MCP, Skills, and CLI
Gate Skills Hub
10K+ Skills
From office tasks to trading, the all-in-one skill hub makes AI even more useful.
Has the AI storage super cycle ended? Supply and demand changes behind the pullback of Micron, Seagate, and SanDisk.
July 3, 2026, the crypto market saw a broad rally. According to Gate data, Bitcoin (BTC) was quoted at $61,506, up 1.7% in 24 hours; Ethereum (ETH) was at $1,698, up 4.7% in 24 hours. Bitcoin rebounded from yesterday's low of $59,776 to $61,507, a gain of 2.86%; Ethereum surged from $1,605 to $1,706, a single-day jump of 6.26%. The US added only 57,000 nonfarm payrolls in June, with the unemployment rate at 4.2%, both below market expectations, easing concerns about a near-term Fed rate hike. Bitcoin's market cap is about $1.23 trillion, and the total cryptocurrency market cap is about $2.21 trillion.
However, traditional financial markets painted a different picture. US stock storage and semiconductor sectors experienced a sharp sell-off, with the Philadelphia Semiconductor Index dropping 11% over two days. The three major storage giants—SanDisk (SNDK), Micron (MU), and Seagate (STX)—collectively plunged, forming a stark contrast to the crypto market's recovery.
What Happened in the Market?
On July 3, Beijing time, the three major US stock indexes were mixed: the Dow rose 1.1% to a new closing high, the Nasdaq fell 0.8%, and the S&P 500 edged down. The storage sector was the biggest drag of the day.
SanDisk (SNDK) suffered the worst decline, closing down 14.13% at $1,745. The stock had already shown weakness the previous trading day, with a cumulative two-day drop of nearly 25%. Micron Technology (MU) fell 5.49% to close at $975.56. In the previous trading session, Micron had already plunged 10.57%. Seagate Technology (STX) fell 10.38%.
Other storage-related stocks also came under pressure: Western Digital (WDC) fell about 10%, Intel (INTC) fell 5%, KLA Corporation fell over 11%, and Arm fell over 6%. In the Japanese and Korean markets, Samsung Electronics fell 6.52%, SK Hynix fell 7.73%, and Kioxia Holdings fell 12.57%. The AI storage sector almost completely collapsed.
Why the Decline?
The sell-off was triggered by a confluence of multiple factors, but the core logic is not a deterioration in fundamentals, but rather a sharp adjustment in market sentiment and trading structure.
First, doubts about the return on AI infrastructure investment intensified. On July 1, news emerged that Meta plans to sell excess AI computing power to external customers. Meta had previously announced that its capital expenditure in 2026 would reach $145 billion. This news was interpreted by the market as: even top tech companies are reassessing the input-output ratio of AI infrastructure. Industry analysts believe that it is not that AI computing power is truly in oversupply, but that cloud service providers must make trade-offs to maintain positive cash flow when computing power investments cannot be fully converted into revenue. However, if AI investments continue to fail to generate revenue, future capital expenditure may be constrained.
The Bank for International Settlements recently also warned about overheated AI investments. Five hyperscale cloud service providers, including Google's parent company Alphabet, Microsoft, and Amazon, plan to invest over $1 trillion in AI-related businesses from 2025 to 2026. The BIS cautioned that while AI investments can improve production efficiency, once the overinvestment recedes and reverses, it could trigger chaos in the financial system.
Second, the previous gains were too large, and technical indicators were severely overheated. Since 2026, AI storage stocks have seen impressive gains. Micron has risen about 240% year-to-date, setting 37 new all-time closing highs in the first half of the year. SanDisk rose 388.4% in the full year of 2025. Seagate surged about 146.32% in the second quarter alone. Such a steep upward slope means that any marginal negative news could trigger massive profit-taking.
Third, short selling became a catalyst for emotional release. According to market news, "Big Short" prototype Michael Burry shorted Micron at around $1,051.87 on or around July 1, becoming the latest catalyst for the continuous sharp decline in AI storage stocks. This news further amplified market panic.
Fourth, the DRAM antitrust lawsuit added uncertainty. On June 25, Samsung Electronics, SK Hynix, and Micron were hit with a class action lawsuit in the US Federal Court in California, accused of colluding to reduce traditional DRAM capacity under the guise of the HBM transition, causing memory prices to surge about 700% over the past four years. The lawsuit seeks triple antitrust damages and an injunction. Although the final outcome of such lawsuits takes time, their suppression of market sentiment is immediate.
Have Fundamentals Deteriorated?
The conclusion is: they have not deteriorated.
Looking at the latest earnings reports, the operating data of the three companies remain strong. Micron's fiscal third quarter of 2026 (ended May 28) reported revenue of $41.46 billion, far exceeding the $9.3 billion in the same period last year and analysts' expectations of $35.91 billion; adjusted EPS of $25.11, above the expected $20.86. The company expects capital expenditure of about $10 billion in the fourth fiscal quarter and about $27 billion for the full fiscal year; capital expenditure in each quarter of fiscal 2027 will be higher than the level of the fourth fiscal quarter of 2026.
SanDisk's fiscal third quarter of 2026 (ended April 30) reported revenue of $5.95 billion, up 97% quarter-over-quarter and 251% year-over-year, far exceeding the company's own guidance and Wall Street's expectation of $4.7 billion; adjusted EPS of $23.41, compared to analysts' expectation of only $14.51. The company expects fourth fiscal quarter revenue of $7.75 billion to $8.25 billion and adjusted EPS of $30 to $33.
Seagate's fiscal third quarter of 2026 reported revenue of $3.11 billion, non-GAAP EPS of $4.10, and free cash flow of $953 million. Fourth-quarter revenue guidance of $3.45 billion and adjusted EPS of $5.00 both exceeded analysts' expectations.
From a supply-demand fundamental perspective, the shortage pattern has not changed.
TrendForce data shows that in the second quarter of 2026, traditional DRAM contract prices are expected to rise 58% to 63% quarter-over-quarter, and NAND Flash contract prices are expected to rise 70% to 75% quarter-over-quarter. Entering the third quarter, the overall DRAM landscape remains extremely tight.
Goldman Sachs expects that on a calendar year basis, NAND average selling prices will surge significantly in 2026—up about 4.5 times year-over-year—and continue to grow by 38% in 2027, higher than the previous expectation of 27%. Goldman Sachs' channel checks show that major storage manufacturers still prioritize capital expenditure toward DRAM rather than significantly increasing new NAND capacity; under the backdrop of expanding AI demand, new NAND supply increments may not be significantly released until 2028.
Morgan Stanley's updated supply-demand model shows a 15% global NAND supply gap in 2026 and a 9% gap in 2027. AI-related NAND demand is expected to increase from 205 EB in 2025 to 400 EB in 2026, further climbing to 609 EB in 2027, with an annual growth rate of about 60%. AI demand as a share of the overall NAND market will rise from 18% in 2025 to 32% in 2026, reaching 41% in 2027.
UBS expects DRAM supply shortages to persist until the second quarter of 2028, and NAND supply shortages to persist until the fourth quarter of 2027. Server SSD demand will grow 56% in 2026 and 47% in 2027.
For HBM, TrendForce expects HBM shipments to grow 60% year-over-year in 2026 and another 60% in 2027. HBM will remain in short supply until 2027. HBM wafer input as a share of total DRAM wafer input will rise from 18% in 2025 to about 30% in 2027.
The Real Contradiction in the Industry: Short-term Trading Crowdedness vs. Long-term Supply Shortage
The core contradiction in the current market is not "whether AI storage demand is real," but "whether the short-term trading structure is sustainable."
From a long-term perspective, the construction of AI infrastructure is still in its early stages. According to Sigmaintell data, global AI infrastructure and computing power investment will maintain double-digit growth annually from 2024 to 2028, with year-over-year growth still at 51% in 2026, although slowing from 104% in 2025, it remains in a high-growth phase. AI server storage consumption is 4 to 5 times that of general-purpose servers. A top-tier AI training server alone carries over 100 GB of HBM, with storage cost accounting for over 40% of the total, a 25-percentage-point increase compared to 15% for traditional servers.
However, there are indeed significant risks in the short-term trading structure. The gains in these stocks have far exceeded the improvement in fundamentals. While Micron's 240% year-to-date gain and SanDisk's 251% revenue growth are impressive, the slope of stock price appreciation is even steeper. When the market begins to question the sustainability of trillion-dollar capital expenditures by hyperscale cloud service providers, any marginal change could trigger a sharp correction.
What Should Investors Focus On?
HBM (Focus on Micron and SK Hynix). HBM is the most critical storage component for AI training and inference, remaining in short supply until 2027. As one of the three major HBM suppliers, Micron directly benefits from this structural trend. HBM consumes about 4 to 5 times the wafer volume of DDR5, meaning capacity expansion is naturally constrained.
NAND (Focus on SanDisk). SanDisk holds a significant position in the NAND Flash field, with enterprise SSDs driven by AI inference becoming the largest application market for NAND. Goldman Sachs expects NAND prices to continue rising sharply in 2026 and 2027. The global NAND supply shortage is expected to persist until 2027.
HDD Data Storage (Focus on Seagate). Cold data storage demand from AI data centers is reshaping the HDD market landscape. Western Digital's and Seagate's hard drive capacity for 2026 has been largely pre-booked by AI data centers. After surging about 146% in the second quarter, Seagate has pulled back, but its fundamental support remains.
Conclusion
On July 3, 2026, the collective decline of the three AI storage giants was a classic "good fundamentals, bad price" moment. Fundamentals—the DRAM and NAND supply-demand gap, the structural shortage of HBM, and the long-term capital expenditure of AI data centers—have not materially deteriorated. What has really changed is the market's repricing of "AI investment returns" and profit-taking on the astonishing gains of the past year.
The crypto market saw a broad rally on the same day, with BTC climbing above $61,500 and ETH breaking through $1,700. The two markets appear to be diverging, but they share the same macro backdrop: marginal improvement in liquidity expectations and easing rate concerns from nonfarm data. However, the AI storage sector in TradFi is still digesting its own unique valuation pressure.
For investors, the key is to distinguish between "short-term trading noise" and "long-term structural trends." The long-term demand logic for AI storage hasn't changed—the supply-demand gaps for HBM, NAND, and enterprise SSDs are unlikely to be fundamentally reversed before 2027. But the process of squeezing short-term valuation bubbles may not be over. Morningstar research director Lorraine Tan said in a Bloomberg TV interview that AI-related stocks may need to fall another 20% to 30% before becoming attractive for purchase, citing new supply from Samsung and SK Hynix and a possible plateau in AI capital expenditure.
The storage industry is standing at the crossroads of a super cycle. The direction of the cycle hasn't changed, but the turbulence is intensifying.
FAQ
Q1: Is AI memory chip supply really in excess?
Currently, there is no substantial supply excess. TrendForce data shows that in the third quarter of 2026, DRAM remains in a "severely tight" state. Morgan Stanley's model shows a 15% global NAND supply gap in 2026. The current sell-off mainly reflects market concerns about "possible future excess," not current reality.
Q2: Why did SanDisk's stock price fall the most?
SanDisk rose 388.4% in the full year of 2025, making it one of the biggest gainers in the storage sector. The large cumulative gains meant that profit-taking pressure was also the most concentrated. Additionally, the NAND market is perceived as having slightly less certainty than DRAM in institutional eyes, amplifying the sell-off magnitude.
Q3: Have Micron's fundamentals been problematic?
No. Micron's fiscal third quarter of 2026 revenue was $41.46 billion, with EPS of $25.11, both far exceeding expectations. The company expects capital expenditure of about $27 billion in fiscal 2026, with further increases in 2027. The decline is more due to valuation digestion after a 240% year-to-date gain and short selling.
Q4: Is the long-term investment logic for AI storage still valid?
The long-term logic remains unchanged. AI server storage consumption is 4 to 5 times that of general-purpose servers. Goldman Sachs expects NAND prices to rise 4.5 times in 2026 and another 38% in 2027. UBS expects DRAM supply shortages to persist until the second quarter of 2028. The key issue is the timing of valuation alignment with fundamentals.