Author: Gelong
Overnight, U.S. stock AI industry chain stocks plunged sharply, with the Nasdaq closing down 1.33% and the Philadelphia Semiconductor Index dropping nearly 5%. Stocks related to storage, optical communication, and AI computing power were generally hit hard.
Panic sentiment quickly spread across markets. Today, Asia-Pacific markets opened with full declines. South Korean stocks Samsung Electronics and SK Hynix tumbled over 6%, while leveraged ETFs such as the double-long Samsung and double-long SK Hynix funds plummeted over 14%, with leveraged fund stampedes further amplifying the volatility.
A-shares' AI industry chain also saw a significant impact. Core sectors such as computing power, optical modules, and storage generally fell over 5%, with several leading companies in sub-sectors dropping more than 8%.
The negative stimuli primarily stemmed from several aspects.
First, the sudden escalation of U.S.-Iran tensions pushed up oil prices and inflation expectations, suppressing the valuation of the entire growth sector from a macro perspective.
The U.S. announced the suspension of negotiation contacts with Iran, further intensifying regional conflicts. Risks in the Strait of Hormuz heated up, driving international oil prices higher to a three-week high. Simultaneously, the UAE has announced the suspension of all trade and financial dealings with Iran, further tightening regional tensions.
Rising oil prices directly fuel concerns about an inflation rebound, prompting the market to reassess the Federal Reserve's monetary policy room.
Long-term U.S. Treasury yields surged sharply, with the 30-year yield briefly touching its highest level since 2007, and the 10-year U.S. Treasury yield also rising significantly.
Under a high-interest-rate environment, the high-valuation, high-capital-expenditure AI sector is hit most directly.
AI computing power construction heavily relies on debt financing. This year's AI-related bond supply has far exceeded previous annual expectations. The yield on bonds issued by Blackstone for Microsoft's data centers is already approaching junk bond levels, directly reflecting the rising financing costs for AI infrastructure.
Goldman Sachs noted that massive AI capital expenditures combined with sovereign deficits are funneling large amounts of capital into the bond market, potentially forcing the Fed to maintain a tight policy stance even in a weakening economic data environment.
As financing costs continue to rise, the market is beginning to re-evaluate the logic of expanding computing power at all costs, leading to a sell-off in the AI hardware sector first.
Secondly, significant divergence has emerged in the commercialization of AI large models, with OpenAI's disappointing performance shattering the market's linear optimistic imagination for the AI application end.
Newly disclosed Q2 data shows OpenAI's quarterly revenue grew only 18% sequentially, with operating losses continuing to widen. Concurrently, high-level executives have left one after another, raising market doubts about the company's internal management stability.
Although competitor Anthropic achieved explosive revenue growth and slight profitability, differences in revenue recognition exist between the two, and this does not mean the entire industry has smoothly entered a profitable era.
OpenAI's decelerating growth has made the market realize that the commercialization of AI large models is not a smooth journey, with huge challenges remaining in enterprise customer conversion and cost control.
Previously, the market was accustomed to unconditionally believing the AI capital expenditure story. Now, investors are beginning to question how much real revenue and profit the massive computing power investment can actually translate into. The AI industry chain has officially entered a phase of assessing commercialization capabilities, moving away from the 'burn cash to scale' stage.
Third, the ongoing U.S.-South Korea semiconductor investment dispute continues to ferment, exacerbating uncertainty in the global memory industry chain and directly impacting the HBM memory sector, a core component of AI computing power.
South Korea has publicly denied related reports about U.S. demands to prioritize building memory chip factories in the U.S. South Korea itself has already planned a massive fund exceeding 580 billion dollars to invest in domestic chip and data center clusters.
If Samsung and SK Hynix are forced to establish large-scale memory production lines in the U.S., it would consume significant corporate capital and weaken the domestic semiconductor industry ecosystem.
However, U.S. pressure is multi-dimensional, using not only tariffs as leverage, but also implying that delays in investment commitments could spill over and affect U.S.-South Korea security cooperation.
This is the reality of 'earning money from the U.S. market, but keeping the capital in the U.S. and complying with industrial demands.' Korean memory companies reap huge profits by supplying HBM to the U.S. AI market, while the U.S. demands these companies repatriate capital to build factories locally or face trade penalties.
The market worries that if subsequent negotiations remain protracted, whether South Korea chooses compromise or confrontation, it will disturb the global memory supply structure.
If South Korea compromises, corporate capital will be diverted, and profits will be eroded by high U.S. construction costs. If it takes a tough stance, it may face trade barrier risks.
This dilemma directly triggered fund selling in Samsung and SK Hynix, with leveraged ETFs further amplifying the decline. Panic sentiment spread outward along the storage industry chain.
Of course, a short-term sharp decline does not mean the AI industry logic has completely ended.
Long-term demand for AI computing power objectively exists, but the market is no longer willing to pay a high premium for an indefinitely optimistic long-term narrative.
Moving forward, market focus will shift from merely looking at the scale of capital expenditure to examining corporate actual profitability, changes in financing costs, and the ultimate direction of global supply chain games.
For the A-share market, external sentiment shocks mostly bring short-term disturbance. Subsequent observation needs to focus on the domestic industry chain's own orders and profit realization, distinguishing between short-term emotional sell-offs and fundamental deterioration.
