Investors have closely seen the GPU Technology Conference (GTC) of a week of NVIDIA for news and updates of the dominant chips manufacturer that feed artificial intelligence applications.
The event arrives at a crucial moment for Nvidia actions. After two years of monster profits, the action has dropped 15% in the last month and 22% below the maximum of January of all time.
As part of the event, CEO Jensen Huang asked analysts questions about topics ranging from the demand for his advanced Blackwell to the impact of Trump administration rates. Here is a breakdown of how Huang answered, and what analysts had been asked, during some of the most important questions:
Blackwell’s demand
Huang said that “he surrendered” the demand on a slide that showed 3.6 million in estimated Blackwell shipments to the four main cloud service providers this year. While Huang recognized speculation about demand reduction, he said that the amount of calculation necessary for AI has “exploited” and that the four largest cloud services clients remain “completely invested.”
Morgan Stanley analyst Joseph Moore said that Huang’s comment on Blackwell’s demand in data centers was the first dissemination in such a way.
“It was clear that the reason why the company made the decision to give that data was to re -make the narrative about the strength of the demand profile, since they continue to present questions related to the open expenditure related to the AI that change from 1 of the 4 to another of the 4, or the pressure of the ASICs, which come from these 4 clients,” Moore wrote to customers, referring to the specific integrated circuits of the application.
Piper Sandler analyst Harsh Kumar, said the slide was “just scratching the surface” on request. Beyond the four largest clients, he said that others are also “all online that he seeks to have both calculator and their budgets allow it.”
Physics
Another conclusion for Moore was the growth of physical AI, which refers to the use of technology for the actions of food machines in the real world instead of the software.
In previous GTCs, Moore said physical AI “felt a bit like speculative fiction.” But this year, “now we are listening to developers fighting with tangible problems in the physical field.”
Meanwhile, the true analyst William Stein described the physical AI as something that is “starting to materialize.” The next wave for physical AI focuses on robotics, he said, and presents a possible $ 50 billion market of $ 50 for Nvidia.
Stein Highliugheted Jensen The demonstration of Isaac Gr00t N1, a customizable base model for humanoid robots.
Macroeconomic concerns
Several analysts highlighted Huang’s explanation of what tariffs for the Nvidia business mean.
“Management said they have been preparing for such scenarios and are beginning to manufacture more on land,” said Da Davidson’s analyst Gil Luria. “It was mentioned that Nvidia is already using [Taiwan Semiconductor’s’] Arizona Fab where it is manufacturing production silicon. “
Bernstein’s analyst Stacy Rasgon said that Huang’s response made Nvidia’s impulse to relocate some manufacturing to the United States would limit the effect of the highest rates.
Rasgon also pointed out that Huang ruled out the concerns of a recession that harms customer spending. Huang argued that companies would first reduce spending in the areas of their business that are not growing, said Rasgon.
