AI trading makes a comeback, Nasdaq hits new high, storage sector leads gains.

AI trading made a strong comeback, coupled with continued upward revisions to earnings expectations, and the Nasdaq 100 index closed at a record high on Tuesday. The memory chip sector led the gains, with NAND flash memory accelerating its transformation from a commodity to a core component of AI infrastructure.

Meta's Muse AI model quickly topped the Apple App Store's free app charts after its launch, igniting optimistic expectations for AI demand. The Nasdaq 100 closed up 0.8% on Tuesday, hitting a record high and breaking its previous high from early June during the session ; the Nasdaq Composite also hit new highs for the second consecutive day, with Monday's single-day gain of 2.3%, its biggest in seven weeks. Meanwhile, the S&P 500 closed slightly lower, highlighting the market's divergence.

The storage sector emerged as the biggest winner in this rally. Sandisk surged 6.8% in a single day, Seagate Technology rose 4.8%, and Western Digital climbed 3.7%. Falling oil prices and reports of progress in US-China AI negotiations further fueled the upward momentum of tech stocks.

FOMO sentiment reignites, tech stocks "all return to the market".

The core driver of this rebound is the market's repricing of AI trading. Emmanuel Cau, Barclays' chief European equity strategist, stated that "AI FOMO (fear of missing out) is making a strong comeback," noting that the rally is driven by "multiple factors," including falling oil prices and reports related to US-China AI negotiations. He also emphasized that after the deleveraging in June, the portfolio structure of the technology sector is "significantly cleaner," and this rally is based on this foundation.

Michael Zigmont, co-head of trading at Visdom Investment Group, described Monday's trading as "a moment when everyone jumps back into the pool," believing the current move "looks like the start of a new momentum rally for tech stocks."

Data from Bespoke Investment Group shows that when the Nasdaq Composite Index rises by more than 2% in a single day and simultaneously hits a new high, the average return over the following month has historically reached 3.7%, with returns over three months, six months, and one year also exceeding the average. Tuesday's Nasdaq close up 0.5%, further confirming this historical pattern.

Upward revision of earnings expectations opens a window for valuation reassessment.

The current market rally is supported not only by sentiment but also by fundamentals. Mona Mahajan, head of investment strategy at Edward Jones, said that AI trading "is likely to continue to have momentum next year," driven by huge capital expenditures on AI infrastructure, while the widespread penetration of AI in productivity will further broaden the scope of beneficiaries.

In its research report, Yardeni Research noted that "investors have noticed that analysts are raising earnings forecasts faster than stock prices are rising," which could prompt the market to reassess the valuations of some large-cap technology and semiconductor stocks.

Mahajan also stated that the Fed's rate hike this month is more of a "mid-cycle adjustment" in policy rates than the start of an aggressive tightening cycle. Coupled with the Atlanta Fed's GDPNow model's forecast of 5.1% annualized real GDP growth for the third quarter (a significant jump from 1.5% in the second quarter), the macroeconomic foundation for corporate profit expansion remains solid. The upcoming third-quarter earnings season is expected to be the next catalyst.

NAND Repricing: Storage Moving Towards a Key Component

The strong performance of the storage sector reflects the market's renewed recognition of the strategic value of NAND flash memory. Rosenblatt analyst Kevin Cassidy released a research report on Monday, initiating coverage of SanDisk with a buy rating. His core argument is that AI systems are "creating opportunities," repositioning NAND from a traditional commodity storage product into a "more systemically critical component in AI infrastructure."

Cassidy points out that traditional NAND demand is driven by increased density and decreased unit cost, but large AI models and inference workloads "increasingly prioritize density, performance, durability, and supply certainty over the lowest absolute price." This shift in demand structure has fundamentally changed the pricing logic of NAND.

In terms of technology, SanDisk's BiCS8 and BiCS10 architectures, jointly developed with Kioxia of Japan, achieve higher density through vertically stacked storage cells. Cassidy believes this will support SanDisk in maintaining its "dual advantage in performance and cost" in the large-scale AI storage market. He predicts that SanDisk's data center revenue will reach $21.7 billion in fiscal year 2027 and further increase to $28.6 billion in fiscal year 2028, and gives a target price of $2,400, implying an upside of about 30% from Tuesday's closing price of $1,887.04.

Cassidy also specifically mentioned that SanDisk has signed multi-year commercial agreements with eight major NAND customers, allowing them to secure supply by locking in prices. He believes that these agreements cover most of SanDisk's capacity and "are expected to improve demand visibility and reduce the volatility of historical NAND cycles," thereby providing more stable support for the company's valuation.

Risk Warning and DisclaimerInvesting involves risk; please exercise caution. This article does not constitute personal investment advice and does not take into account the specific investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their specific circumstances. Any investment decisions made based on this information are at your own risk.