Nvidia’s strong earnings report revealed that the company is still wearing the crown of AI chips. So why did the stock fall in Thursday trading? According to CNBC’s Jim Cramer, that’s because investors are worried that NVIDIA’s biggest customers won’t be able to keep spending at such a breakneck pace unless AI efforts start to significantly increase profits. Nvidia stock fell about 4% on Thursday to about $188 a share, retreating toward a flat line from the start of the year. “There’s a perception that all we really care about, or should care about, is how much money Amazon, Microsoft and Meta owe, and are they ultimately not going to make money?” Cramer said Thursday on “Squawk on the Street.” Cramer’s Charitable Trust, a portfolio used by CNBC Investing Club, owned Nvidia stock long before OpenAI’s ChatGPT sparked an AI boom in late 2022. “This is skepticism about their customers, not skepticism about Nvidia,” Cramer emphasized. Amazon, Microsoft, Metaplatform and Google’s parent company Alphabet (often referred to as a hyperscaler) reported stronger-than-expected capital spending in their latest earnings reports. Total spending between the two companies this year is expected to be about $700 billion. However, such aggressive plans are putting pressure on free cash flow. Free cash flow is the money a company has left over after paying expenses and purchasing equipment, which it uses to pay down debt, pay dividends, buy back stock, and invest in growth. Amazon is expected to have negative free cash flow in 2026, according to Wall Street estimates compiled by FactSet. Alphabet is expected to report sales of $26 billion, down 64% from a year ago. Metaz is modeled for an 86% decline to $6.1 billion. Microsoft is a relative outlier. Analysts expect free cash flow to be $71.3 billion for the fiscal year ending in June, about the same as a year ago, according to FactSet. During Wednesday night’s earnings call, NVIDIA CEO Jensen Huang was asked whether he was confident customers could continue to increase capital spending. “I’m confident that their cash flow will increase,” he said. “The reason is very simple. We are now seeing the use of agent AI and the usefulness of agents in every enterprise in the world. Because of that, we are seeing an incredible demand for computing. In this new AI world, computing is revenue.” “We are generating profitable tokens that are productive for customers and profitable for cloud service providers,” Huang said. Tokens are the fundamental building blocks of AI computing. “To generate tokens, you need computing power, and that translates directly to growth, which translates directly to revenue,” the CEO added. But Kramer said the hyperscaler’s bondholders, who have borrowed money to buy Nvidia’s chips, don’t want to hear about future returns. They want to hear about profits. Cramer acknowledged that there is some truth to their claims, but stressed that he remains confident in the feasibility of building AI and Nvidia’s role in it. This is especially true since Nvidia is seeing a lot of growth from other types of customers. That’s what CFO Colette Kress said Wednesday night. While hyperscalers accounted for just over half of Nvidia’s data center revenue in the fourth quarter of fiscal 2026, Kress said the company’s remaining data center customers drove growth “as we diversify our revenue.” While the hyperscaler debate may not disappear anytime soon, Cramer said he’s excited about Nvidia’s growing relationships with Claude creators Anthropic, OpenAI, and others. Nvidia and Anthropic announced a technology partnership in November. He said OpenAI and Anthropic are “just blossoming into huge customers.” Nvidia chips trained OpenAI’s ChatGPT. Kramer said the key takeaways from Thursday’s stock price are: “Customers have no choice. They have to spend money. And you can say, maybe they don’t have enough money. Well, there’s going to be a gap where there’s income but no profit. But they’re all afraid of not spending.”
