Wall Street’s enthusiasm for artificial intelligence took a hit this week as investors grew more uneasy about the enormous sums technology companies are pouring into the buildout. Shares of major AI beneficiaries sold off as the market began to question whether the spending spree will generate enough profit and productivity to justify the cost.
The scale of that investment is immense. Alphabet, Amazon, Meta Platforms and Microsoft together plan to spend up to $720 billion this year, with most of the money aimed at AI-focused data centers. Supporters of AI argue that the technology will transform the global economy, but critics have increasingly warned that the rush of money into the sector may be creating a bubble.
Those concerns showed up quickly in trading. Amazon and Alphabet each dropped about 5% on Monday. On Tuesday, the selling spread to companies supplying the hardware needed for the data-center expansion, with Nvidia, Micron Technology, Broadcom and Lam Research among the stocks leading the broader market lower.
Much of the early AI push was financed with internal cash by Microsoft, Alphabet and other hyperscalers. Now, however, those companies are leaning more heavily on capital markets to fund their expansion plans, a sign that the cost of competing in AI is rising fast.
Alphabet, Google’s parent company, said earlier this month that it is raising $80 billion by selling stock to help cover its investment plans. The company expects to spend as much as $190 billion this year, a sum greater than the entire market value of The Walt Disney Co. Alphabet has also said its investment spending next year will “significantly increase.”
Amazon took a similar approach in March, selling $54 billion in bonds across the U.S. and Europe as it prepares to invest around $200 billion this year in AI. Elon Musk’s rocket company SpaceX, meanwhile, had been in a three-day slide going into Tuesday. It recovered some ground but still finished the day slightly below its closing price from its first day of trading on June 12. Musk has said SpaceX will need to spend heavily to carry out plans to place AI data centers in space, and the company has announced that part of a coming bond sale will be used for its AI expansion.
Chipmakers have been major winners so far because AI data centers and related projects have sharply increased demand for memory chips and processing power, creating shortages and driving prices higher. Investors have pushed up those stocks on expectations of large future profits, though traditional valuation measures suggest some may now be expensive.
Marvell Technologies is one example. After losing money for five consecutive years, it posted a profit of $2.7 billion for the fiscal year that ended in January, helped by growth in its data-center business. Its stock has risen more than threefold this year, and its price-to-earnings ratio has climbed from roughly 30 at the beginning of 2026 to nearly 100.
Some data-storage names have surged even more dramatically. Sandisk shares have jumped more than 700% so far this year, leaving the stock with a P/E ratio of 68. Whether that valuation proves excessive may depend on the company’s ability to meet aggressive Wall Street forecasts over the next year: analysts expect earnings per share of $188.05, versus $29.16 per share for the 12 months ended March 31. Using that projected figure, Sandisk’s P/E falls to about 11. For comparison, the S&P 500 currently trades at around 25 times earnings.
Investors on Tuesday pulled back from some of those highfliers. Sandisk fell 13.6%, while Marvell dropped 9.4%. The retreat also dragged down exchange-traded funds with large technology positions. The Invesco QQQ Trust Series ETF lost 3.3%, and the iShares Semiconductor ETF fell 7.9%.
Not all of the selling necessarily reflected a loss of faith in AI itself. With major indexes having recently notched repeated records, some investors may simply have used the downturn to lock in gains after a powerful run. “With no clear catalyst driving the move lower, we believe today’s pullback likely reflects profit-taking following a strong rally from the March lows,” said Brock Weimer, an investments strategy analyst at Edward Jones.
Technology stocks have been the main engine behind this year’s record-setting market advance. Inside the S&P 500, the tech sector is up nearly 27% over the past three months and about 17% for the year. In Asia, South Korea’s Kospi has nearly doubled in 2026.
Heavy selling in that market on Tuesday prompted a trading halt in the Kospi and helped set the tone for the wave of U.S. tech-stock selling that followed when American markets opened, Wedbush analyst Dan Ives said in a research note. Even so, Ives said AI demand in Asia remains strong, writing that overall enterprise demand is “showing no cracks in the armor, which continue to make us very bullish on owning the tech AI winners over the coming year.”
Others are more cautious about what happens if companies keep racing to expand AI infrastructure at the current pace. Philip Straehl, chief investment officer at Morningstar Wealth, said the surge in spending could eventually create too much capacity. “Periods of elevated capital investment have historically not translated into strong outcomes for investors, leaving us cautious on the outlook,” Straehl wrote in a report last week.
He said the rapid growth in AI computing power could eventually pressure prices, reduce returns and lead to a slowdown in investment. Semiconductor companies, Straehl warned, are “particularly exposed to this dynamic.”