Wall Street Morning News: Philadelphia’s semiconductor sector plunges into a bear market, memory chips collapse, the Fed’s hawkish tone returns, and OpenAI is accused of being the AI version of “Lehman Brothers”

Every Monday to Friday morning, focusing on macro, U.S. equities, AI, precious metals, and crude oil—using data to review the market and using trends to seize opportunities, produced by PANews.

Rate-cut expectations get targeted by hawks, U.S. stocks weigh down

U.S. stocks ended the prior streak of consecutive rebounds. Tech stocks came under broad pressure, and the AI industrial chain became the main direction for profit realization; meanwhile, Federal Reserve officials once again released hawkish signals. The escalating situation in the Middle East boosted risk-hedging sentiment, and global capital returned to a tug-of-war around two main lines: whether high interest rates will be maintained for longer, and whether AI investment returns are sufficient to justify lofty valuations.

All three major U.S. stock indexes closed lower. The Dow Jones Industrial Average fell 0.20%, the S&P 500 dropped 0.51%, and the Nasdaq Composite fell 1.47%. Although U.S. economic data overall still showed resilience, a wave of concentrated selling in the semiconductor sector made the AI theme the core force dragging on the index.

Strong economic data not only failed to lift risk appetite, but also became ammunition for hawkish Fed officials to push down rate-cut expectations. U.S. June retail sales excluding gas stations rose 0.7% month over month; initial jobless claims fell to a low of 208k; the Philadelphia Fed Manufacturing Index surged to 41.4. Morgan Stanley’s wealth management division pointed out that consumers are still continuing to spend and the labor market shows no sign of loosening. Goldman Sachs also quickly raised its tracking estimate for Q2 GDP growth to 2.4%.

More than the economic data, what truly moved the market was the Fed officials’ latest comments.

• Dallas Fed President Logan clearly signaled support for modest rate hikes, warning that the month-over-month CPI decline “is not enough,” and saying that if inflation cannot return to 2% on its own, policy restrictions must be taken.

• Kansas City Fed President Schmid followed suit, reiterating that its core concern remains inflation and opposing ignoring part of the price increases.

• Of even greater importance to the market is that Fed Vice Chair Jefferson publicly discussed for the first time the impact of AI on inflation. He said that if AI infrastructure investment, compute capacity build-out, and expansion in consumer demand grow faster than improvements in production efficiency, then AI could instead become a new source of inflation in the coming years. This means AI investment is not only changing the tech industry—it is also entering the Fed’s policy framework.

Crude oil hangs in the balance at the “$80” lifeline; the dollar rebounds, gold and silver pull back

The Middle East situation remains tense. Iran carried out Phase 11 of its “flash operation,” using drones to attack U.S. forces’ base in Bahrain; the U.S. then increased military strikes against Iran, targeting including bridges near the Strait of Hormuz. Iran warned the U.S. not to touch the Strait of Hormuz, calling it an “uncrossable red line.”

WTI crude rose to around $81 at one point overnight, then quickly retreated and closed down at $79. A trader at CIBC Private Wealth Group said the physical market has not yet sent signals of severe supply shortages, so traders are unwilling to blindly push up oil prices. However, $80 has become the “main switch” for risk assets: once U.S. oil holds that level, the market will inevitably reprice the deadly chain of “higher oil prices—inflation expectations—U.S. Treasury yields—tech stock valuations.”

Precious metals showed a clear correction. Spot gold fell below the $4,000 per ounce integer level, with a daily decline of more than 2%. Spot silver fell below $55 per ounce, hitting the lowest level since late November last year. Institutions such as Fidelity International still maintain a long-term bullish logic, saying central bank gold buying and macro uncertainty support have not changed, but near-term rates and the dollar remain strong headwinds.

Semiconductor stocks lead the decline; semiconductors slip into a technical bear market; the debate over an AI bubble is reignited

What stood out most in the U.S. market overnight was not the index decline, but the structural breakdown in semiconductor stocks. The Philadelphia Semiconductor Index fell 4.29%, more than a 22% pullback from the mid-June peak, officially entering a technical bear market. The Semiconductor ETF fell 3.70%, while the Technology Sector ETF fell 2.24%.

TSMC’s Q2 net profit was 706.6 billion New Taiwan dollars, with a gross margin approaching 68%. It also raised full-year capital expenditures to between $60 billion and $64 billion, and lifted its full-year sales growth forecast to more than 40%. Under traditional logic, this is solid evidence of hot AI demand. But the market flipped to sell, because traders have started worrying that “the higher the AI capital expenditures, the longer the return cycle.”

Memory stocks became the epicenter of the selloff. In South Korea, regulators tightened a single-stock leveraged ETF, directly triggering deleveraging pressure: the minimum margin was raised from 10 million won to 30 million won, margin is limited to cash, a maximum of 20 shares can be bought per trade, and the issuance of new single-stock leveraged products is prohibited. Leveraged capital was forced to withdraw, and highly volatile names such as SK hynix, SanDisk, Seagate, and Western Digital were sold off in a concentrated fashion.

JPMorgan research shows that over the past five to six weeks, hedge funds have sharply cut their AI-related exposures and leveraged ETF holdings. Bloomberg strategist Tatiana Darie said the selloff in chip stocks is nearing the technical threshold of multiple prior “bottoming rebounds” in recent years, but whether it can stabilize depends on whether hyperscale cloud providers continue raising AI capital expenditures.

Meanwhile, overseas markets have seen ongoing intensifying debate over OpenAI’s business model. Ed Zitron, a commentator who has long been bearish on AI, published a long piece saying that the essence of an AI bubble is the “OpenAI bubble.” If OpenAI fails, it could become “Lehman Brothers” for the AI era, hitting data centers, AI infrastructure, and global tech stock valuation. But long-term investors including Howard Marks argue that AI still belongs to a general-purpose technology revolution, and the industry is still in the early stage of commercialization, so it should not be simply viewed as a bubble.

Specific project moves and stock price fluctuations:

• Memory chip sector takes a collective hit: SK hynix plunges 13.69%, SanDisk falls 12.63%, Seagate Technology drops 10%, Western Digital declines 9.22%, Micron falls 5.65%, and market cap drops back below $1 trillion. Tightening of trading in a single-stock leveraged ETF by South Korean regulators is seen as the direct trigger for this round of selling. At the same time, the market worries about too-rapid industry capacity expansion; increased supply in the future will compress profit margins. Funds quickly exit overvalued AI hardware assets, making the memory segment the sub-sector with the largest intraday decline.

• Optical communications continues to adjust: Corning falls more than 9%, Lumentum drops more than 6%. The market is worried that the future growth rate of data center capital expenditures may slow down, which could affect the demand growth rhythm for optical modules and optical communications equipment; funds simultaneously reduce the proportion of AI infrastructure allocations.

• Google down 4.44%: The release of Gemini 3.5 Pro’s flagship model is delayed by several months compared with the original plan due to model performance—especially code ability not yet meeting internal expectations—raising market concerns about Google’s competitive advantage in AI. While the long-term R&D schedule has not changed, the market worries it may fall further behind in competition with OpenAI and Anthropic. Among related large model AI stocks, Meta falls 2.46%, and Amazon drops 1.99%.

• Nvidia down 2.40%. The company launched Cosmos 3 Edge, a world model aimed at robots and vision agents, and plans to form a physical AI industry alliance in Japan. Japan also plans to purchase 27.5k next-generation Rubin chips for local robot AI model development; but overall AI hardware deleveraging outweighed product catalysts. Among related chip stocks, TSMC ADR falls 2.25%, AMD drops 5.33%, and Intel falls 5.84%.

• Apple up 1.76%, hitting new highs again: The OLED iPad mini is expected to be released as early as this autumn, with a launch expected around October. The entry-level iPad and Air series will be updated next year. iPad sales have exceeded Wall Street expectations for two consecutive quarters, and product upgrades strengthen the hardware recovery narrative.

• SpaceX down 3.08%: After an IPO, the first major Starship test flight is canceled. Musk said some engines failed to start, triggering an automatic launch abort procedure. Risk appetite in related newly listed stocks weakened. This year, the weighted average return of IPO companies in the U.S. fell to 6%, lagging the roughly 11% gain of the S&P 500.

• Netflix down more than 8% after hours: Q2 results meet expectations, but Q3 revenue growth guidance hits the lowest level in nearly three years, raising concerns that growth has peaked. The company emphasized it will increase investment in live sports broadcasting, video podcasts, and AI content production, but short-term investors focus more on the slowdown in guidance.

• Oracle down 6.25%: The market continues to worry that AI cloud infrastructure has high capital expenditures. The company has seen a clear cumulative pullback from recent highs. However, Piper Sandler still maintains a “buy” rating with a target price of $225, believing that capital expenditures will be converted into larger-scale cloud compute resources and drive future revenue.

What to watch next

• July 17-20: Shanghai World Artificial Intelligence Conference (WAIC) and the Global Governance High-level Meeting. Focus will be on China’s AI implementation outcomes and how governance proposals may affect global standards, which could boost sentiment in related industrial chains.

• July 18-19: AGI Summit in San Francisco. Developments from OpenAI and other big players will directly influence where U.S. stocks’ AI valuations go.