The Magnificent Seven of the US stock market have shrunk.

Since the start of the second-quarter earnings season, US-listed tech giants have been going through an awkward collective "tribulation". On one hand, Tesla's share price has plummeted, and Alphabet (Google's parent company) saw its stock price drop sharply after releasing its earnings report; on the other hand, the price of Credit Default Swap (CDS), which reflects the debt default risk, has kept rising, and the debt premiums of giants including Oracle, NVIDIA, Meta and even Alphabet have hit record highs one after another.

The stock market is deflating bubbles, and the bond market is sounding alarm bells. Many people are confused: large language models are evolving, cloud computing revenue is growing, and US tech giants still have hundreds of billions of dollars in cash on their books, so why have they suddenly "lost their appeal"?

The "Magnificent Seven" index has fallen by more than 10%

The core of the problem lies in the fact that Wall Street's underlying logic for treating AI has changed. If the market in the past two years was playing an arms race where "the one who spends the most wins", now the market has entered a cruel "reckoning moment".

Let's first take a look at the account books just submitted by the giants.

Although Alphabet's second-quarter revenue exceeded 100 billion US dollars and its cloud business grew rapidly, its quarterly capital expenditure reached more than 40 billion US dollars to build data centers and purchase computing power, which directly led to its free cash flow turning negative rarely in more than 20 years since its listing. The management even raised the upper limit of this year's capital expenditure to more than 200 billion US dollars in one go.

The day after the earnings release, Alphabet's share price plummeted by more than 7%. As of the close of US stock market this Monday, Google's share was quoted at 326.56 US dollars, down more than 20% from the intraday high of 408.37 US dollars set in May this year.

Tesla also paid a price for its rapid expansion in AI, computing power and robotics. Its free cash flow in the second quarter plummeted to negative, its operating profit margin shrank to a dismal 1.4%, and its share price plunged 14% the day after the earnings report. As of the close of US stock market on Monday, Tesla's share was quoted at 309.22 US dollars, down more than 30% from the high of 453.40 US dollars in May.

The other companies among the US "Magnificent Seven" have not yet released their earnings reports. But according to previous forecasts by analysts surveyed by FactSet, Meta is expected to report negative free cash flow when it releases its second-quarter results this week. Analysts expect the same situation to happen to Amazon, which is also due to release its earnings report this week, and its free cash flow already fell into negative territory in the first quarter.

As of the close of US stock market on Monday, Meta's share was quoted at 593.87 US dollars, down more than 13% from the high of 690.88 US dollars in April this year; Amazon's share was quoted at 231.39 US dollars, down nearly 17% from the high of 278.56 US dollars in May; Microsoft's share was quoted at 389.1 US dollars, down more than 16% from the high of 466.32 US dollars in May; NVIDIA's share was quoted at 196.51 US dollars, down nearly 17% from the high of 236.27 US dollars in May.

The only company among the Magnificent Seven whose share price has not shrunk is Apple. As of the close on Monday, its share was quoted at 336.91 US dollars, hitting an all-time intraday high, with a cumulative increase of more than 22% since the beginning of the year, leading the "Big Seven Tech Giants". Apple was previously criticized for insufficient AI investment, but now it seems that it has successfully avoided the capital expenditure trap.

Judging from the US Magnificent Seven Tech Index, as of the close on Monday, it stood at 20732.35 points, down more than 10% from the high of 23117.14 points in May this year.

The bond market is also sounding the alarm

In the stock market, people may still be willing to pay for dreams. But in the bond market, what matters is the confirmed ability to repay principal and interest.

In the past, US tech giants relied on the huge cash flow from their core businesses to "support" AI R&D. But now, the appetite of the cash-devouring AI is so huge that even cash flow behemoths like Alphabet and Meta are beginning to feel stretched, and have to frequently use debt leverage.

Oracle announced last month that it would spend 70 billion US dollars to build data centers in the coming year, and S&P immediately downgraded its credit rating to "BBB-", which is close to junk level.

According to a report by the Financial Times of the UK on the 27th, the cost of Meta's latest debt financing for a 12-billion-US-dollar data center in Texas has risen sharply, with the yield approaching the level of junk bonds, which intuitively reflects that the capital's willingness to allocate AI-related bonds continues to weaken.

John Aylward, Chief Investment Officer of asset management firm Sona, said: "The credit market is extremely averse to uncertainty. At this stage, the investment pace and capital cost of AI financing are completely unpredictable, which has triggered a serious market confidence crisis." Regarding the transaction price of Meta's debt issuance pegged to Class B junk bonds, Aylward commented that "it is very abnormal, but it is also a true portrayal of the current market".

Data from the London Stock Exchange Group (LSEG) shows that the prices of Credit Default Swaps (CDS, a commonly used tool to hedge corporate debt risks) related to companies including Oracle, SpaceX, Alphabet, Amazon, Meta, Broadcom and NVIDIA have recently risen to record highs.

In layman's terms, CDS is a "default insurance" bought for corporate debt. A very low premium generally indicates that the market considers the company to be extremely stable with a very low probability of bankruptcy, and insurers are willing to underwrite for a small profit. A surge in CDS premiums indicates that the market is beginning to worry about the company's debt pressure.

"The biggest question is whether this level of capital expenditure (capex) will grow indefinitely, and when will we see the turning point where cash flow turns positive again?" said David Brown, Global Co-Head of Investment Grade at Neuberger Berman, "We won't get the answer in the short term, which also explains the weak performance."

"This could become a problem, because there is still a large amount of financing queuing up to be completed," Brown said.

This concern is spreading beyond Oracle, and the cost of providing protection for NVIDIA's five-year debt has also reached a record high of 79 basis points.

George Catrambone, Head of Americas Fixed Income at DWS Group, said that although investors still assess the possibility of default by investment-grade issuers to be low, buying CDS has become a way for investors to protect themselves from future credit rating downgrades and market volatility.

"Hedging is becoming more and more necessary, especially after seeing these capital expenditure figures after the earnings releases," Catrambone said. "A large amount of debt has been issued, but it may not generate corresponding revenue. The market is subjecting it to more and more scrutiny."

CDS has also become a broader market alternative indicator for bearish bets on tech stocks.

"For hyperscale cloud vendors, focus on CDS instead of EPS (Earnings Per Share)." said Manish Kabra, Head of US Equity Strategy at Societe Generale, "AI capital expenditure is still outpacing cash generation, pushing the free cash flow of tech groups to a cyclical low."

When cash-rich tech giants start borrowing heavily from the bond market like high-risk, high-leverage enterprises, and their free cash flow starts to bleed, the bond market will inevitably impose "risk penalties" by raising CDS rates and increasing debt issuance costs. And the warning from the bond market, in turn, has poured a basin of cold water on the stock market.

Going forward, tech giants not only need to prove that their large language models are "smart enough", but also need to prove that these infrastructures that have poured in hundreds of billions of dollars in costs can truly deliver a clear and sustainable commercial return model. If they fail to do so, the recent double kill of stocks and bonds is probably just the prelude to this big reshuffle.

The views in this article are for reference only and do not constitute investment advice. Investment is risky, so be cautious when entering the market.

This article is from the WeChat official account "China News View" (ID: jwview), written by Luo Kun, edited by Li Xiaoxuan, chief responsible editors Xue Yufei and Chang Tao, and published by 36Kr with authorization.