The primary engine of the AI trading frenzy appears to be stalling, subjecting the U.S. technology bull market to its most severe stress test of the year.
On Friday, global semiconductor stocks faced continued selling pressure. The Philadelphia Semiconductor Index (SOX) fell as much as 5.7% intraday before closing down about 1.6%, marking a cumulative decline of approximately 20.2% from its record high in late June and entering a technical bear market. Simultaneously, leading chip stocks like NVIDIA (NVDA), Advanced Micro Devices (AMD), and Broadcom (AVGO), which had spearheaded the AI rally, remained under pressure. The market is beginning to reassess whether AI-related capital expenditure can continue to justify current valuations.
In stark contrast to the chip sector, Apple (AAPL) emerged as a new safe haven for capital. Intraday, Apple's market capitalization briefly surpassed NVIDIA's, allowing it to reclaim the title of the world's most valuable publicly traded company. This development signals the most pronounced internal rotation within the U.S. tech sector this year—capital is shifting away from the crowded AI infrastructure trade toward mega-cap platform companies with more stable cash flows and stronger consumer electronics attributes.
Meanwhile, the stock of SpaceX continued its sharp decline, closing below its IPO price for the third consecutive day on Friday and hitting a new low since its public debut over a month ago.
Even a better-than-expected earnings report from TSMC on Thursday failed to halt the broader tech sell-off. An increasing number of analysts are now debating a question rarely taken seriously before: has the nearly two-year AI-driven tech bull market already reached an inflection point?
Semiconductor Index Enters Technical Bear Territory
Over the past week, semiconductors ranked among the worst-performing global sectors.
Data shows the Philadelphia Semiconductor Index fell nearly 9% for the week, marking a decline of over 20% from its June peak, which meets the market's definition of a "technical bear market." The iShares Semiconductor ETF (SOXX), which tracks the industry, also entered bear market territory, posting one of its worst weekly performances since last year.
More significantly, this adjustment has spread from individual stocks to the entire supply chain.
Core AI beneficiaries like NVIDIA, AMD, Arm, Broadcom, and Micron have seen consecutive sharp declines. In Asian markets, companies like Samsung Electronics, SK Hynix, Tokyo Electron, and Advantest simultaneously faced capital outflows, with Japanese and Korean tech indices seeing notably widened losses.
Analysts note market concerns have shifted from "whether AI demand exists" to "whether AI capital expenditure has already priced in growth for several years ahead."
The key narratives that previously drove the rally to new highs—unlimited expansion of AI computing investment, persistent GPU shortages, and cloud giants' continuous increases in capital expenditure—are now being critically re-examined by the market.
Some hedge funds have begun reducing exposure to AI infrastructure-related holdings, with capital flowing more into previously lagging sectors rather than exiting the equity market entirely.
Apple Challenges NVIDIA for Market Leadership
Amid the collective decline in chip stocks, another development in the U.S. market has sparked widespread discussion.
During Friday's trading session, Apple's market cap reached approximately $4.88 trillion, while NVIDIA's fell about 3.5%, reducing its value to around $4.86 trillion. On an intraday basis, this meant Apple briefly regained its position as the world's most valuable listed company, a title it had not held since last April.
Subsequently, as NVIDIA's share price pared some losses, its market cap once again overtook Apple's. At the close, Apple's stock rose 0.14%, setting a new closing record for the third consecutive day, with a market cap of about $4.9 trillion. NVIDIA's stock fell 2.21%, with a market cap of approximately $4.91 trillion. Apple could close the gap and officially reclaim the top spot in the next trading session.
This shift quickly became a hot topic on social platforms. Many investors view it as symbolic of a change in market leadership, where AI infrastructure is no longer the sole anchor for valuation.
Discussions in investment communities suggest the market consensus over the past year that "NVIDIA cannot lose" is weakening. Apple is regaining favor due to its stable cash flows, ecosystem advantages, and the upcoming new AI-powered Siri. Other investors interpret this as capital beginning to prefer large-cap tech stocks with higher earnings certainty over continued bets on high-investment, high-valuation AI hardware companies.
Market observers note that while Apple was previously seen as a laggard in the AI race, this perception is changing as investors reassess its long-term competitiveness in the AI era.
Several Wall Street institutions point out this does not signify a market rejection of AI. Instead, it suggests the AI trade is expanding from the "shovel sellers" to platform companies that can truly commercialize AI and generate cash flow.
Memory Sector Shows Divergence
It is noteworthy that not all chip stocks fell in unison.
Shares of South Korean memory chip leader SK Hynix's ADR moved against the trend, rising intraday. The stock was up about 9.9% at its session high before paring most gains to close up 1.1%, reversing a two-day plunge of over 20% and becoming one of the few large semiconductor companies with a positive performance on Friday.
Market analysis attributes this to three main reasons.
First, High Bandwidth Memory (HBM) remains one of the most constrained components in AI servers, with a supply-demand dynamic significantly more favorable than most other chip categories outside of GPUs.
Second, compared to the substantial gains in the GPU sector over recent months, the memory sector had undergone more thorough adjustments earlier, resulting in relatively less valuation pressure.
Additionally, some capital is rotating within the AI supply chain, moving from higher-valued GPU leaders to memory chip firms with more certain earnings prospects.
However, most institutions view this more as a structural rotation within the sector rather than a signal of a broader semiconductor industry recovery.
High-Valuation AI Assets Face Reassessment
Another closely watched AI-related stock, SpaceX, has also become a significant casualty in the current tech sell-off.
On Friday, SpaceX's stock fell over 5%, marking the fourth consecutive intraday session and third consecutive close below its IPO price, while also hitting a new low since its June 12th debut. Compared to its market capitalization at the closing high on June 16th, SpaceX's value has plummeted by over a trillion dollars. Furthermore, the stock faced additional pressure after hours due to a cancelled launch mission, worsening market sentiment.
As one of the most sought-after newly listed tech companies this year, SpaceX's continued decline is seen by many investors as the latest signal of rapidly diminishing risk appetite for high-valuation growth stocks.
Recent sharp contractions in market value for AI leaders from NVIDIA and Micron to SpaceX, amounting to trillions of dollars in cumulative losses, indicate that the previously overcrowded AI trade is undergoing intense deleveraging.
Testing the Bull Market's Defenses
Compared to previous adjustments, the key difference in this tech stock decline is that "good news can no longer drive share prices higher."
Even after TSMC reported better-than-expected earnings and raised its capital expenditure forecast, global tech stocks continued their slide. Similarly, strong earnings from Samsung Electronics and SK Hynix earlier failed to prevent the broader semiconductor sector's correction.
This signifies a shift in market trading logic.
Over the past year, AI capital expenditure could seemingly override all valuation concerns. Now, investors are focusing more on several critical questions: How long can the hyperscale cloud providers sustain their capital expenditure? When will AI infrastructure investments translate into commercial returns? Is the current profit growth sufficient to support valuations after a period of significant expansion?
Based on market performance, the answers remain unclear.
However, capital is not exiting the tech sector entirely but is instead being reallocated internally. On one side, high-beta assets like semiconductors and AI hardware are seeing continued reductions. On the other, cash-rich, mature business model tech giants like Apple are regaining investor favor.
For Wall Street, this may indicate that the AI bull market is not over, but the "anchor" driving market gains is changing—shifting from "who sells more GPUs" toward "who can first turn AI into stable profits." The most resilient line of defense in this tech rally is being tested, and new market leaders may already be emerging.
Comments