Summary: On August 6, US AI hardware stocks staged a dramatic reversal: storage chips plummeted during the session before soaring straight up, while optical communication stocks collectively surged. SanDisk and Western Data delivered "explosive" earnings reports but were initially sold off due to guidance below sky-high expectations. This article provides an in-depth analysis of the surge in AI storage demand and the fulfillment logic of tech giants' capital expenditures, outlining the layout opportuniti
On the morning of August 7 Beijing time, the US stock market staged a breath-taking reversal. At the open, artificial intelligence (AI) hardware stocks collectively plummeted, with the storage chip sector "crashing," and the Philadelphia Semiconductor Index falling over 2% at one point; however, just hours later, bottom-fishing capital rushed in, storage chip stocks soared straight up, and optical communication stocks surged, with the Philadelphia Semiconductor Index反而收涨逾1%. In the same period, the Dow Jones opened high and closed low, plunging over 400 points, ending its previous five-day winning streak. Between this fire and ice contrast, investment opportunities in the AI hardware industrial chain are being re-priced by the market.
"Explosive" Earnings First Sold Off: The Unexpected Start of Storage Giants
The trigger for this major reversal came from the "explosive" earnings reports just released by storage chip giants—SanDisk and Western Data.
SanDisk's earnings showed quarterly revenue up 372% year-over-year to $8.97 billion, far exceeding market expectations of $8.6 billion; adjusted EPS reached $39.25, about 135 times the $0.29 from a year ago; adjusted gross margin reached 84.6%, 3.2 times that of the same period last year. Western Data was equally impressive: quarterly revenue $3.747 billion, up 44% year-over-year; adjusted EPS $3.56, doubled year-over-year; non-GAAP gross margin rose to 54.4%, an increase of about 13 percentage points from the same period last year. The data together point to one fact—AI-driven data storage demand remains extremely strong.
However, the market's reaction was unexpected: after the earnings report, both companies' stock prices fell in after-hours trading, with Western Data plunging over 10% and SanDisk down nearly 5%. The next day, the storage sector sold off across the board, with Western Data down nearly 20% intraday, SanDisk down over 12%, SK Hynix down over 8%, and Micron Technology down nearly 6%.
Expectation "Ceiling" Effect: Good Earnings Are No Longer Enough
Why did such impressive earnings trigger selling? The answer lies in "earnings guidance." SanDisk expects next quarter's adjusted EPS to be $44 to $46, with revenue $10.3 to $10.8 billion; the midpoint of the range is below analysts' expectations of $10.82 billion. Western Data's guidance was roughly in line with expectations but failed to bring additional excitement.
JJ Kinnahan, head of retail business development and alternative investment products at the Chicago Board Options Exchange, summed it up: "When guidance is weak, even if profits and revenue beat expectations, it may not be enough to drive the stock price up." He also noted that profit-taking factors cannot be ignored—over the past 12 months, SanDisk has risen 3000% cumulatively, and Western Data has risen over 530%. In the face of such huge gains, any signal below expectations could become a reason for capital to cash out and leave.
This phenomenon reflects the core characteristics of the US stock market's efficiency: stock prices not only reflect past performance but also digest investors' high expectations in advance. When "surprises" become "in-line with expectations," the market's pricing logic quickly becomes more demanding.
Intraday Major Reversal: Bottom-Fishing Capital and Optical Communication Sector "Take Over"
Just as the market thought the AI hardware trend would die down, a dramatic scene unfolded. On August 6 intraday, after a sharp drop, the storage chip sector encountered strong buying interest, with stock prices soaring straight up and recovering all losses; semiconductor and optical communication sectors led gains, with the Philadelphia Semiconductor Index recovering from its initial 2% decline and rising over 1%.
Optical communication concept stocks had already been strengthening consecutively. Against the backdrop of surging demand for high-speed optical modules in AI data centers, Applied Optoelectronics once rose 19% in a single day, Coherent rose over 12%, Lumentum rose nearly 9%, upstream fiber material supplier Corning also jumped nearly 9%, and Marvell Technology, Fabrinet, Astera Labs and other industrial chain companies simultaneously rose. The alternating activity of storage and optical communication vividly reflects the depth and breadth of the US AI hardware industrial chain—when one sector pulls back, capital quickly shifts to other links in the industrial chain, giving investors more diversified layout options.
AI Capital Expenditures Enter "Fulfillment Period," How to Layout Industrial Chain Opportunities
From a broader perspective, the underlying logic of this AI hardware trend is the accelerated fulfillment of tech giants' capital expenditures. SpaceX's first quarterly report after listing showed its Q2 capital expenditures surged six-fold year-over-year to $18.4 billion, mostly invested in AI business, with Musk clearly stating that future AI computing infrastructure will only use NVIDIA chips; Microsoft, Amazon, Google's parent Alphabet and other cloud giants saw their combined market value surge by nearly $1.5 trillion last week. The continued increase in capital expenditures is being transmitted layer by layer along the chain of "computing power chips—storage—optical interconnects—servers," providing investors with a complete and diversified opportunity map.
For ordinary investors, seizing this round of opportunities can focus on several directions:
- Focus on细分龙头: Pay attention to high-quality companies in storage, optical modules and other细分领域, choosing those with real orders, continuously improving gross margins and cash flow support, rather than simply chasing concepts;
- Make good use of index tools: Diversify individual stock risks by tracking the Philadelphia Semiconductor Index, Nasdaq 100 Index and other ETFs to share the overall growth benefits of the AI hardware industrial chain at a lower threshold;
- Pay attention to valuation and expectation matching: Against the backdrop of huge sector gains and high expectations, avoid chasing highs, and pay attention to marginal changes in earnings guidance and valuation safety margins.
Risk Warning: Clarity Under High Expectations and High Volatility
Of course, opportunities and risks always go hand in hand. Currently, the AI hardware sector's valuation already implies extremely high growth expectations, and marginal changes in earnings guidance may trigger sharp fluctuations; Federal Reserve Governor Lisa Cook warned this week that inflation remains sticky and is ready to support rate hikes, making the interest rate environment's uncertainty不容小觑; the twists and turns of Hormuz Strait navigation talks and repeated geopolitical situations may also disturb global risk appetite. "The Big Short" prototype Michael Burry even bluntly stated that this round of upward self-reinforcing cycle may ultimately end in a sharp adjustment.
The uniqueness of the US stock market lies in its efficient price discovery, abundant liquidity, and diversified industrial structure, providing global investors with full opportunities and choices. Facing the "major reversal" of AI hardware stocks, rational positioning, position control, and respect for expectations may be the correct posture to navigate high volatility cycles.
