US chip stocks plunge, evaporating over $1 trillion in market value! Inflation exceeds expectations, reigniting rate hike prospects; AI infrastructure financing bottlenecks trigger valuation reassessment

According to Zhihu Finance APP, the US semiconductor sector is undergoing a profound market revaluation. With the September inflation data surpassing expectations and further solidifying Federal Reserve rate hike prospects—combined with rare warnings from Broadcom (AVGO.US) and Intel (INTC.US) CEOs about AI infrastructure funding bottlenecks—the global AI-themed trading, centered on semiconductors, has suffered massive sell-offs in the past week.

The Philadelphia Semiconductor Index (SOX) dropped about 12% cumulatively over the past six trading days, wiping out approximately $1.6 trillion in market value, marking the longest losing streak since May 2025. On Wednesday alone, the index plummeted 4.30%. This decline not only affected the US market but rapidly spread to major semiconductor hubs like South Korea and Japan, sparking widespread discussions among investors regarding the sustainability of the AI investment cycle.

During this period, more than half of the index’s constituent stocks fell by at least 10%. Arm Holdings Plc (ARM.US) plummeted by 21%. Applied Materials (AMAT.US) dropped 16%, and Intel (INTC.US) declined 13%. GlobalFoundries (GFS.US) was the only stock to rise, outperforming its peers during the period.

The semiconductor sector’s downturn echoes the broader market weakness—a tech-heavy Nasdaq 100 Index fell by over 2%, marking its worst single-day performance since late July, while the S&P 500 Index fell 1.1%.

This round of sell-offs is not limited to the US. Global semiconductor stocks suffered a synchronous sell-off: in Thursday’s Asian early trading, the Korea Composite Stock Price Index (KOSPI) opened down more than 5%, hitting a new low since July, and the Nikkei 225 plunged more than 3%. The interconnected global semiconductor stock declines reflect investors’ deepening systematic concerns about the sector.

"In recent days, pressure on global chip stocks has intensified," wrote Ulrike Hoffmann-Burchardi, Chief Investment Office at UBS, in a research note on Wednesday. "Although increasing competition in AI, rising funding needs, and higher rates may periodically exacerbate volatility, we believe that this latest sell-off is more of a recalibration of expectations rather than a fundamental break in the AI investment logic."

Inflation pressures heat up across the board, core CPI rebounds more than expected

The latest bout of pressure emerged after September inflation data exceeded expectations, further strengthening market expectations for the Federal Reserve to continue raising rates—even if policymakers may remain on hold at this month’s upcoming meeting.

Data released by the US Department of Labor on October 14 showed that, driven by rising energy and food costs, the September Consumer Price Index (CPI) rose 0.6% month-over-month, marking the largest increase since April this year. The year-over-year increase reached 3.7%, higher than August’s 3.4%, and exceeding economists’ forecast of 3.6%. Excluding the more volatile food and energy prices, core CPI rose 0.3% month-over-month, matching August, but exceeding market expectations of 0.2%. Its year-over-year rise increased from 2.4% in August to 2.6%, also above the expected 2.5%.

Specifically, energy prices were the main driver of overall inflation. In September, the energy price index rose 5.2% month-over-month; gasoline prices jumped 9.1%, contributing more than half of the month’s CPI increase. Ongoing Middle East conflicts continue to push up fuel costs—so far in 2023, energy prices are up 20.4%, with rising fuel costs further pushing up airfare, which rose 1.2% month-over-month in September and is up 22.4% for the year.

Notably, price increases have extended beyond the energy sector and are spreading more broadly. Analysts point out that prices for entertainment, used cars and trucks, education, airfares, healthcare, and apparel have all increased. While some of these increases stem from tariffs and energy factors, others indicate that inflation is proving more stubborn.

Olu Sonola, head of US economic research at Fitch Ratings, noted that the latest inflation data serves as a "red light" for the Federal Reserve, as price increases spreading into broader service sectors could increase the risk of sustained inflation.

A deeper issue is that construction of AI infrastructure is itself becoming an inflation driver. The data center construction boom is pushing up prices of semiconductors and other goods, with labor costs in construction channels further transmitting these pressures across the broader economy.

Chicago Fed President Goolsbee explicitly pointed out that the inflation pressures spurred by the AI boom are expanding beyond directly related industries: “The areas of price increases are no longer limited to data center components and industries directly connected. I am increasingly concerned this is a typical case of overall demand overheating.”

Inflation data that surpassed expectations rapidly reshaped market expectations for the Fed’s interest rate path. According to the CME "FedWatch" tool, after the CPI data was released, the odds of a 25-basis-point Fed hike in October at one point jumped from around 18% pre-release to 50%. The odds for a hike in December climbed to nearly 90%. While these expectations have cooled somewhat as of press time, the probability of a December hike remains above 60%.

The Fed’s decision-makers issued a flurry of hawkish signals before and after the inflation data release. New York Fed President Williams said he still believes one more rate hike this year is appropriate, but does not see a need for the central bank to raise rates further in 2027. Williams noted that the Iran conflict pushing up energy prices and the widening refined oil price spread changed his short-term inflation outlook, making him "firmly support" the September rate hike decision. Boston Fed President Collins also stated she expects “modestly further rate hikes will be appropriate” before the end of the year to ensure inflation returns in a timely and sustained manner to the Fed’s 2% target.

The Cleveland Fed President expressed a similar stance, believing that further monetary policy tightening may be needed. Citi adjusted its forecast following the CPI data, now expecting the Fed to hike 25 basis points in December 2026, whereas previously it had forecast rates to remain unchanged this year.

In an interview, Williams especially highlighted that the AI infrastructure frenzy, along with tariffs and energy prices, has become one of the three core variables affecting the US inflation path. He observed an emerging upward trend in goods and costs associated with AI construction: “We are indeed seeing higher prices for goods and other costs connected to AI buildout: data centers, enormous demand for chips. There have been quite sudden and substantial price increases in AI-driven sub-sectors, and we are seeing some spillover into construction labor costs.”

However, he also provided a relatively optimistic long-term outlook—that as chip capacity expands and productivity improves, supply could improve in the future, exerting a moderating force on inflation.

Broadcom CEO’s warning triggers chain reaction: Capital constraints on AI investment come to light

Beyond macro-level rate pressures, industry-level funding constraint warnings became another key catalyst for this round of sell-offs. Earlier this week, Broadcom CEO Hock Tan likened some AI-chasing investors to “gamblers,” frankly stating that many AI investors have little comprehension of the technologies they are investing in, yet hold sky-high expectations. He further warned that capital—not chip supply—will become the greatest constraint on AI development.

Tan also questioned the actual benefits of enterprise AI spending, emphasizing that AI model developers must find clients who can generate returns on investment via AI, rather than relying on “test phase users who are simply wasting money on Tokens.” He illustrated this with a vivid analogy: “It’s like when you went to the arcade as a kid and your parents gave you $10 worth of tokens, and you spent them all in 10 minutes and learned nothing. That’s exactly what’s happening in the corporate world now.”

Tan also questioned the actual benefits of enterprise AI spending, emphasizing that AI model developers must find clients who can generate returns on investment via AI, rather than relying on “test phase users who are simply wasting money on Tokens.” He illustrated this with a vivid analogy: “It’s like when you went to the arcade as a kid and your parents gave you $10 worth of tokens, and you spent them all in 10 minutes and learned nothing. That’s exactly what’s happening in the corporate world now.”

His comments sparked both widespread resonance and controversy in the market. Michael Burry, famed investor and real-life inspiration for “The Big Short,” reposted Tan’s remarks on X, underscoring the issue that companies spending on AI Tokens do not necessarily generate real value.

Broadcom’s warning is supported by real-world evidence. AI cloud computing provider CoreWeave (CRWV.US) saw its 2032 unsecured bonds plunge below 80 cents on the dollar on Wednesday, with the yield exceeding 15%. Its credit default swaps rose to record highs, with market pricing indicating a more than 50% probability of default in the next five years. CoreWeave’s shares closed more than 15% lower on Wednesday.

D.A. Davidson analysts previously dubbed CoreWeave the “subprime tranche” of the AI sector, warning that if the AI funding boom fades, the company may be unable to service its hefty debts.

Daniel O'Regan, Executive Director of Equity Trading at Mizuho Securities, noted that in the past weeks, investors have grown “increasingly unwilling to fund AI-related spending at any cost.” Capital expenditures for AI have outstripped companies' abilities to generate free cash flow, meaning companies must seek external financing to buy necessary hardware. As external funding grows scarcer, spending on chips and other AI infrastructure could be materially impacted.

A team of J.P. Morgan analysts led by Nikolaos Panigirtzoglou wrote in a report on October 14 that although current stock movements are similar to the pullback in late July, the background for AI infrastructure stocks is now more unfavorable.

Panigirtzoglou noted the reason is “Although overall equity funds are still seeing inflows, retail investors have so far failed to ‘buy the dip’, buyback flows have fallen off a cliff, short-base is insufficient, and leveraged equity ETF assets under management remain higher.” “Despite the recent pullback, from the perspective of positioning or fund flows, cyclicals don’t seem to be in oversold territory.”

“We advise investors to focus on which areas of AI expenditure are transforming into sustainable earnings growth, as well as broader enterprise and consumer adoption,” added UBS’s Hoffmann-Burchardi.