Hyperscalers spend so much money on AI that they can’t beat investors

AI For Business


Investors were quick to express their displeasure when Alphabet raised its 2026 AI spending forecast by $15 billion to $190 billion to $205 billion in its earnings report last Wednesday. Alphabet stock fell as much as 7.4% in Thursday trading.

It was the latest sign that the market is increasingly doubtful that the huge sums of money hyperscalers are spending on AI development will soon pay off.

Investors worry they are watching a repeat of the dot-com bubble, when companies spent heavily on internet infrastructure but saw poor initial profits, leading to a collapse in stock prices. They recently cheered Meta on reports that the company plans to lease excess cloud capacity and turn all its demand into cash, sending the company’s stock soaring 21% in a matter of days.

So should hyperscalers follow the collective wisdom of the market and slow down their spending plans?

Last Thursday, as Alphabet’s stock was plummeting, I posed a question to Molly Pieroni, president of Yachtman Asset Management, which invests in Alphabet.

Her answer was a resounding “no.” This is an interesting contrast to the prevailing narrative at the moment, which comes from people actually involved in the game.

“We always value discipline, but in this case, you really have to compete to get on the other side where you want to be,” Pieroni said. “So we’re not saying they need to stop investing in capital equipment and reduce their investments.”

We highlight current Catch-22 dynamic face-to-face hyperscalers. If they cut back on spending, investors may reward them in the short term. But in the long run, it could upset investors like Pieroni, who think all the spending is justified, and lose out in the AI ​​race.

In a sense, investors are right to demand financial discipline from the companies they own. However, they may also be overreacting to some extent to any expense. Pieroni pointed out that Google, in particular, has a healthy balance sheet and a fairly diversified business that it can rely on to improve its bottom line. After all, the company beat its second-quarter profit.

For now, most hyperscalers appear to continue to ignore investor protests over their spending. We’ll learn more about their thinking later this week when Microsoft, Meta, and Amazon report earnings.

Last week, I also spoke with Brandon Nelson, a top-performing fund manager at Calamos Investments. He continues to bet on hyperscaler capex beneficiaries even as investors have dumped them in recent weeks over fears that spending will stop. He said the current market jitters are likely to be temporary in the ongoing bull market and that overspending will continue.

“From 1996 to the peak of the stock market in 2000, there were 10 corrections in the 10% range in the Nasdaq. Those were not pleasant ones,” he said. “My guess is that there is a gut check like that in building AI infrastructure.”