Wall Street's fear of an AI slowdown is 'laughable,' Morgan Stanley says

One of Wall Street's top banks isn't worried about the artificial intelligence trade even as it struggles amid fears of lower investment by companies in 2025.

"The idea that we are in a digestion phase for AI is laughable given the obvious need for more inference chips which is driving a wave of very strong demand," Morgan Stanley analyst Joseph Moore said in a note on Friday.

Moore pointed to recent comments from OpenAI's Sam Altman and Alphabet's Sundar Pichai as evidence that AI companies still can't get enough GPU chips.

"While Wall Street wrings its hands over a laundry list of very real concerns, Silicon Valley focus has shifted to a very different challenge - growth in tokens generated of more than 5x since the beginning of the year is very much stressing the ecosystem, and is driving a surge in investment to handle those workloads," Moore explained.

AI stocks have been particularly weak this year, a trend sparked by the late-January debut of DeepSeek's efficient large language model. The model sparked fears that cloud hyperscalers might need fewer GPU chips from Nvidia to develop their AI capabilities.

The AI bubble fears grew after President Donald Trump unleashed a wave of tariffs in early April.

Shares of AI darling Nvidia have plunged 28% since late January, and mega-cap tech giants closely tied to the AI trade are down about 21% since their peak just a few months ago.

Turning to Nvidia, Moore admitted that supply constraints and export restrictions on its H20 chips would likely limit the company's revenue upside over the next few quarters. Still, once supply constraints are worked out, they should see a significant leg of growth in 2026.

"NVIDIA had almost no revenue for Blackwell in October, did $11 bn in January, and likely well over $30 bn in the current quarter," Moore said, adding that he doesn't expect the growth to slow down. anytime soon.

"Per our checks, this demand commentary has intensified in the last few days," Moore said. "We have been highlighting this strong inference demand recently but it continues to get stronger."

Moore increased his calendar year 2026 revenue and earnings per share estimates for Nvidia by 10.7% and 11.9%, respectively, adding that the chip giant remains a "top pick."

The analyst reiterated his "Overweight" rating on Nvidia and set a $160 price target, representing potential upside of 45% from current levels.