Google, Tesla Post-Earnings Slump Sends Nasdaq Down 2%: Why SK Hynix, Micron and Other Memory Stocks Still Buck the Trend?

TradingKey - On July 23, Eastern Time, the three major U.S. stock indexes fell across the board, with Google ( GOOGL) and Tesla ( TSLA) releasing their second-quarter financial results, where both saw free cash flow turn negative, and massive capital expenditures put pressure on their stock prices. However, performance diverged at the sector level, with memory chip stocks bucking the trend to rise today, and the Philadelphia Semiconductor Index edged up 0.09%.

As of press time, the Dow Jones Industrial Average fell 1.02% to 51,683.54 points; the Nasdaq Composite Index fell 1.95% to 25,189.19 points; and the S&P 500 Index fell 1.07% to 7,418.67 points.

[Three Major Stock Indexes Trend Chart, Source: FutuBull]

At the sector level, mega-cap tech stocks fell across the board, but AI hardware stocks that benefit from rising capital expenditures bucked the trend to post gains.

Among cloud providers: Google (GOOGL) fell 6.79%, Amazon ( AMZN) fell 4.14%, Meta Platforms ( META) fell 3.40%, Microsoft ( MSFT) fell 1.64%.

Among memory chip stocks: SK Hynix ADR ( SKHY) rose 6.3%, Micron ( MU) rose 3.3%, SanDisk ( SNDK) rose 1.9%

The driver behind the stock price divergence between upstream and downstream sectors of the AI industry chain is the transmission pathway of capital expenditure. The market believes that the massive AI capital expenditures of Google and Tesla are precisely the core logic driving the counter-trend rise of AI hardware stocks.

A consensus had previously formed in the market: the deployment of AI servers is inseparable from the support of high-bandwidth memory (HBM), DRAM, NAND flash memory, and storage devices. However, there was always disagreement over whether upstream giants, such as cloud providers, would continue to ramp up capital expenditures.

The financial reports from Google and Tesla completely broke this uncertainty—the massive capital expenditures of both companies are, in essence, advance bets on AI computing power infrastructure. This investment will eventually be transmitted step-by-step along the industry chain to upstream segments such as memory chips and server hardware. Financial data showed that Google's second-quarter capital expenditure reached $44.9 billion, nearly doubling from the same period last year, with the vast majority allocated to servers, data centers, and network infrastructure construction.

Against the current backdrop of tight memory chip supply and scarce computing resources, the greater the capital expenditure, the more robust the orders for hardware suppliers will be—which is precisely the underlying logic behind the divergence of 'tech stocks falling, AI hardware stocks rising'. As hyperscale cloud providers continue to expand data centers, the market expects memory chip makers to directly benefit from stronger pricing and sustained high demand.

Andrew Slimmon of Morgan Stanley said on Wednesday that memory chips and computing power will remain in short supply for the foreseeable future, reinforcing his positive outlook on AI-related investments, despite ongoing market skepticism over massive data center capital expenditures.

Slimmon said that Wall Street currently remains focused on the short-term investment costs of AI infrastructure, while tech companies are investing aggressively to capture long-term demand. "Memory chips are scarce, and computing power is scarce," Slimmon said bluntly. "I don't think this shortage will be resolved anytime soon."

He further suggested that investors should focus on areas of scarcity, believing that the supply side has not yet caught up with the continuously growing demand for AI infrastructure. Although he admitted that some memory chip stocks showed signs of overheating earlier this year, the recent correction has made valuations more attractive.

At the same time, Slimmon refuted comparisons of the current AI capital expenditure cycle to the dot-com bubble or the crash phase of a commodity cycle, pointing out that the end of the dot-com bubble occurred only after supply caught up with demand—and current industry conditions indicate that the wave of AI infrastructure construction is still in its early stages.