Nvidia stock stems losses as GTC leaves Wall Street analysts 'comfortable with long term AI demand'

Updated Wed, Mar 19, 2025, 6:36 AM 3 min read

Nvidia (NVDA) stock rose more than 1% Wednesday morning, stemming losses from a two-day slide that saw shares drop 5% as the AI chipmaker’s annual GTC event failed to excite investors amid a broader market downturn.

Nvidia stock’s reversal comes as Wall Street analysts walked away from CEO Jensen Huang’s closely watched GTC 2025 keynote Tuesday optimistic about the company’s roadmap and AI demand, doubling down on their bullish outlooks on the chipmaker in notes to investors. Those outlooks counter broader concerns about long-term AI demand and more efficient AI models reducing the need for computing hardware such as Nvidia’s acclaimed GPUs (graphics processing units, or AI chips).

“[W]e came out of the keynote reassured in NVIDIA's leadership which if anything seems to be expanding,” Citi (C) analyst Atif Malik wrote, reiterating his Buy rating on Nvidia stock and $163 price target and calling the chipmaker “king of the hill.”

Read more: How does Nvidia make money?

Huang laid out Nvidia’s upcoming AI chips during his presentation in San Jose, Calif., on Tuesday afternoon: Nvidia will launch its upcoming AI chip, Blackwell Ultra, in the second half of 2025; its next AI superchip, Vera Rubin in the second half of 2026; and the next-gen superchip after that (Vera Rubin Ultra) in the second half of 2027. Huang reiterated that he sees data center spending on compute hardware (i.e., Nvidia’s total addressable market) reaching $1 trillion.

“I've said before that I expect data center build-out to reach $1 trillion. And I am fairly certain we're going to reach that very soon,” he told a large crowd.

Raymond James (RJF) analyst Srini Pajjuri reiterated his Strong Buy rating on Nvidia, echoing Citi’s Malik: “Overall, we walked away comfortable with long term AI demand and continue to be impressed with NVDA’s roadmap & technology innovation.”

Bernstein’s Stacy Rasgon wrote in his own note to investors Wednesday morning, “The roadmap looks really solid, and their capability gap vs competitors across their entire massive stack continues to widen.”

Huang also seemed to convince analysts that cheaper, cost-efficient AI models (e.g., that of China’s DeepSeek) won’t hurt AI chip demand.

“During the keynote session, Jensen Huang presented a compelling case on how innovations in AI models, such as DeepSeek, will drive higher compute complexity and necessitate increasing compute demand,” wrote JPMorgan (JPM) analyst Harlan Sur in a note early Wednesday.

Some were more skeptical.

“There was little commentary comparing the use cases [of Nvidia's GPUs] vs. peers or ASICs [custom chips made by companies such as Broadcom seen as competition for Nvidia's GPUs],” Jefferies (JEF) analyst Blayne Curtis wrote late Tuesday, adding that Vera Rubin “will only be an incremental update in 2026 with Rubin Ultra the more meaningful leap ahead in 2027.”