Nvidia CEO Jensen Huang’s feedback on his firm’s This fall earnings name on Wednesday might in the future be remembered as the height of the AI bubble—the basic second that happens in each bubble when hubris and self-delusion overtake widespread sense.
For that to not be the case, it might imply that, starting in 2026, the U.S. launched into one of many best and most unprecedented financial expansions in historical past.
It’s a state of affairs that Huang clearly believes in. His message to traders on Wednesday: Massive Tech’s huge spending on AI know-how, significantly Nvidia’s chips, just isn’t wherever close to completed. “This new way of doing computing is not going to go back,” he stated, and companies are “going to be building out this capacity from this point forward and continue to expand from here.”
Nvidia delivered completely blockbuster leads to the ultimate three months of 2025, as demand for its AI chips went by means of the roof. Income elevated an astounding 73% to $68.1 billion, and Nvidia stated gross sales within the present quarter would broaden by as a lot as 200%.
If Nvidia’s inventory was up lower than 1% after these heroic outcomes, it’s as a result of there’s a basic downside at play. Greater than half of Nvidia’s income comes from the 5 huge “hyperscalers”—that’s, the Googles and Amazons of the world (Nvidia didn’t explicitly title the 5 firms, nevertheless it’s simple sufficient to guess who they’re), who’re feverishly shopping for as lots of Nvidia’s GPU chips as they will to stuff within the huge AI knowledge facilities they’re setting up.
Many of those hyperscalers have vowed to double their capital expenditures this yr as they construct extra knowledge facilities. Meta, which spent $72 billion on capex in 2025, plans to spend as much as $135 billion this yr. Google stated it can spend as a lot as $185 billion, in comparison with $91 billion the yr earlier than. All informed, the large hyperscalers are budgeting almost $700 billion in capex this yr.
The plain query is: How lengthy can this go on? These hyperscalers are already outspending their prodigious free money movement and elevating debt to finance the AI infrastructure buildout. If that group of 5 firms doubles capex yearly, we’re $2.8 trillion of spending by 2028, and $5.6 trillion by 2029.
The Wall Road analysts on Wednesday’s earnings calls requested Huang about this. How sustainable is that this, actually? Will the opposite 50% of Nvidia’s clients assist maintain the AI infrastructure spending spree going? What sort of functions and real-world makes use of will drive demand for all this new AI infrastructure?
Huang walked by means of the logic for perpetual spending as calmly and confidently as a professor explains a basic math downside to a scholar.
“If you think about it and said ‘OK, well the world was investing about $300 to $400 billion a year in classical computing, and now AI is here and the amount of necessary computation is 1,000 times higher… if we continue to believe there’s value in it, then the world will invest to produce that token,” Huang stated, referring to the fundamental unit of knowledge processed by AI fashions.
“So the amount of token generation capability that the world needs is a lot more than $700 billion,” he continued. “And I’m fairly confident that we’re going to continue to generate tokens, we’re going to continue to invest in compute capacity from this point out.”
When it comes to functions, the latest buzz round AI brokers and instruments like Open Claw is already creating a brand new wave of demand. “Agentic AI has reached an inflection point, and it literally happened in the last 2 or 3 months,” Huang stated. After agentic AI, he added, there will probably be bodily AI, as new AI fashions are built-in into robotics and manufacturing gear.
“AI is here. AI is not going to go back. AI is only going to get better from here,” Huang stated.
In different phrases, the celebration is simply getting began and the music just isn’t about to cease. No less than to not Huang’s ears.
