It’s not a bubble, Nvidia CEO says as company crushes Wall Street expectations again

Nvidia CEO Jensen Huang firmly dismissed mounting investor anxieties over an "AI bubble" this week, arguing that the company's continuous, record-breaking financial results are proof of a fundamental, structural transformation in global computing.

The statement came after the AI chip giant yet again crushed Wall Street expectations for the third quarter of fiscal year 2026. Nvidia reported record revenue of $57.0 billion, a staggering 62% jump year-over-year, and issued stronger-than-expected guidance for the current quarter, forecasting $65.0 billion in revenue.

The bubble talk vs the numbers

For months, analysts have warned that the massive capital expenditure by cloud giants, including Amazon, Google, and Microsoft, on AI infrastructure could be unsustainable, raising comparisons to past tech boom-and-bust cycles. Huang, whose company is the sole bottleneck supplier of the chips fueling this boom, directly addressed the fear on the earnings call.

"There's been a lot of talk about an AI bubble," Huang stated. "From our vantage point, we see something very different. We are unlike any other accelerator; we excel at every phase of AI."

The numbers backed his assertion: Data Center revenue soared to $51.2 billion, up 66% year-over-year, and company executives noted that sales of the new Blackwell GPU platform are "off the charts," with cloud GPU capacity "sold out."

The three transitions: Why demand is real

Huang laid out a three-part case to analysts, emphasizing that demand is driven by deep, non-speculative industry shifts that require Nvidia's Accelerated Computing platform:

• From CPU to GPU computing: As Moore’s Law slows, traditional CPU-based systems cannot handle the heavy computational demands of AI. Industries are moving to GPUs to achieve better performance per dollar across workloads like data processing, search, and engineering simulation.

• The rise of generative AI: AI is no longer just improving old applications but creating entirely new software categories. From sophisticated generative models reshaping search to advanced coding assistants, all of which require continuous GPU power for both training and inference.

• Agentic and physical AI: The emergence of next-generation AI systems capable of reasoning, planning, and acting (like robotics and autonomous systems) will require an unprecedented, sustained buildout of specialized computing power.

Huang concluded that Nvidia’s unified architecture supports all three transitions, making the current boom an infrastructure buildout that will generate "trillions of dollars" in value over the next decade, not a fleeting speculative craze.