Technology companies are investing millions to incorporate artificial intelligence into their businesses and build huge data centers. Investors who jumped on the bandwagon appear to be having second thoughts.
Proponents of artificial intelligence believe that it is the next great revolution for the global economy. Revolutions aren’t cheap. Just four companies – Alphabet, Amazon, Metaplatforms, and Microsoft – plan to spend up to $720 billion this year, primarily on AI data centers.
This week, investors are taking note of the huge spending and questioning whether AI will be able to generate returns and productivity worth all the investment. Critics are discussing the possibility of a bubble in AI investing. Amazon and Alphabet fell about 5% on Monday.
On Tuesday, several companies that make the chips needed to build data centers led the market decline: Nvidia, Micron Technology, Broadcom and Lam Research.
Initially, Microsoft, Alphabet, and other so-called hyperscalers relied on cash reserves to fund their AI expansion. But they are increasingly relying on the market to raise cash.
Google’s parent company Alphabet announced earlier this month that it would sell stock to raise $80 billion in cash to fund investments. All told, Alphabet plans to spend $190 billion this year, more than the value of all of Walt Disney Co.’s stock, and Alphabet expects investment spending to increase “significantly” next year.
In March, Amazon sold $54 billion in bonds in the U.S. and Europe, with plans to spend about $200 billion on AI investments.
Elon Musk’s rocket maker SpaceX had been in the doldrums for three days heading into Tuesday. Although the stock regained some of its lost ground, it remained on a small upward trend compared to the closing price on the first day of trading on June 12th. Musk has acknowledged that SpaceX will need to spend a lot of money to accomplish its plans to send an AI data center into space, and the company announced that a portion of its upcoming bond issue will fund its AI buildout.
Many semiconductor companies are soaring as demand for memory and processing power for AI data centers and other projects has led to soaring prices. If investors push up stock prices in hopes of expanding profits, the price-to-earnings ratio, an important indicator of how expensive stock prices are, may soar.
Marvell Technologies lost money for five straight years before posting a $2.7 billion profit in the fiscal year that ended in January, thanks to gains from its data center business. The company’s stock price has more than tripled so far this year, and its P/E ratio has increased from about 30x at the beginning of 2026 to nearly 100x. Some data storage companies are making even more impressive profits. SanDisk stock has soared more than 700% since the beginning of the year, trading at a P/E ratio of 68.
The current price-to-earnings ratio for the S&P 500 is approximately 25.
On Tuesday, investors sold at least some of their holdings in these stocks. SanDisk fell 12.2% and Marvell fell 8.1%.
Exchange-traded funds (ETFs), which invest heavily in tech stocks, also suffered a sharp decline. The Invesco QQQ Trust Series ETF fell 2.6%, and the iShares Semiconductor ETF fell 7.1%.
While some investors may be questioning whether companies that are going full-throttle with their AI infrastructure investments will ultimately be able to generate a return on their investment, some of this week’s selling may be investors holding back to capture some of the gains following the stock market’s recent streak of record highs.
“There is no clear catalyst for a decline, and today’s pullback likely reflects profit-taking following the strong rebound from the March lows,” said Brock Weimer, investment strategy analyst at Edward Jones.
Rising gains in big tech stocks have helped drive record gains in major stock indexes this year. The tech sector alone in the S&P 500 is up about 27% in the past three months alone and about 18% for the year. In Asia, South Korea’s Kospi will almost double by 2026.
Wedbush analyst Dan Ives said in a research note Tuesday that Tuesday’s heavy selling triggered a trading halt in the Kospi and triggered a wave of tech stock selling at the opening of U.S. trading.
Demand for overall AI companies in Asia “sees no cracks in the armor, so we remain very bullish on acquiring technology AI winners next year,” he added.
Still, the race by tech companies to invest in expanding AI infrastructure could ultimately be planting the seeds for future oversupply, said Philippe Strahl, chief investment officer at Morningstar Wealth.
“We are cautious about the outlook because periods of high capital spending have historically not produced strong outcomes for investors,” Strahl said in a note last week.
He expects the rapid expansion of AI computing power to weigh on prices, hurting corporate profits and ultimately leading to a pullback in investment. Semiconductor companies are “particularly exposed to this dynamic,” Strahl wrote.
Short supply and higher prices for computer memory have helped boost stocks like SanDisk this year, but the dynamics of supply and demand will likely cause other companies like Nvidia to look for ways to capture a piece of that market, he noted.
Strahl suggested that as AI-related companies become more represented in major stock indexes, investors could benefit from diversifying into sectors such as healthcare or other areas that are less dependent on AI expectations.
Copyright 2026 Associated Press. Unauthorized reproduction is prohibited. This material may not be published, broadcast, rewritten or redistributed without permission.
