Julie Coleman @ITSJULIECOLEMAN

KEY POINTS

• CNBC’s Jim Cramer on Friday argued against a “new and negative AI narrative” some on Wall Street have adopted that suggests large-scale spending on artificial intelligence has peaked.
• “I believe that AI infrastructure remains the greatest story of our time,” he said. “The data center, the hardware, and the software for it are so much stronger than the bears think.”

CNBC’s Jim Cramer on Friday argued against a “new and negative AI narrative” some on Wall Street have adopted that suggests large-scale spending on artificial intelligence has peaked.

“I believe that AI infrastructure remains the greatest story of our time,” he said. “The data center, the hardware, and the software for it are so much stronger than the bears think.”

AI behemoth Nvidia and semiconductor outfit Broadcom saw their shares fall after their recent quarters beat estimates but failed to blow them out of the water. Cramer said investors read this “relative underperformance” to mean that hyperscalers have been spending too much on the new AI technology.

But Cramer said Nvidia’s gross margin was hurt because the company currently doesn’t currently have enough supply to meet the demand for its new AI chip, Blackwell. He added that spending for new AI hardware can be uneven depending on the customer and said the two companies are “unwitting casualties” of the widespread AI pessimism.

He used a Google versus Bing analogy, saying Microsoft did not spend enough on search, giving Alphabet a long-lasting leg up in the business. Hyperscalers don’t want to fall behind in the same way, he said, suggesting, for example, Meta AI needs chips from Nvidia to complete with programs like Gemini, Clause or ChatGPT. Cramer also noted that Broadcom should benefit from its recent acquisition of VMware and the release of Apple’s newest iPhone.

“In the end, people will realize that there’s no AI bubble, there’s just intense demand and very little supply,” Cramer said. “The Magnificent Seven have had repeated periods of underperformance, but they tend to always come back, as they should. They’re all incredibly well-run, and they have both accelerated computing and artificial intelligence.”