The bubble question everyone's been asking finally got answered by the person who's been fueling the AI frenzy—and his response suggests that investors worried about a 2000-style crash might be missing something fundamental about what's actually happening in computing right now.

Nvidia Corp. (NVDA) CEO Jensen Huang tackled the AI bubble speculation during a panel at the 2025 U.S.-Saudi Investment Forum last week. His three-minute answer spent less time on market valuations and more on technological transformations that are already underway, backed by trillions in global capital commitments.

Moore's Law Hit a Wall—And That Changes the Game

Huang's argument begins with a hard reality in computer science: Moore's Law, that beautiful decades-long trend of doubling transistor density every two years, has plateaued. This isn't a temporary hiccup—it's a structural constraint creating a widening chasm between exploding demand for computing power and the industry's ability to deliver it through traditional chip architectures.

"This isn't speculation—it's a fundamental challenge driving the need for new computing paradigms," Huang explained.

That gap is why companies aren't just experimenting with AI infrastructure for fun—they're pouring massive capital expenditures into it because they have to, not because of hype.

The Supercomputer Revolution That Already Happened

"The world is voting with real capex," Huang said, describing how industries are investing heavily in GPUs because traditional CPUs simply can't handle AI training costs efficiently—a problem that would otherwise require trillions in annual spending to solve the old way.

This isn't speculation about future technology. It's a present-day reallocation of computing resources that's already underway across finance, healthcare, research, and manufacturing. The money is real, the infrastructure is deployed, and it's generating returns today.

Three Waves of Computing—And We're Just Starting the Third

Huang laid out three distinct computing revolutions, each building on the previous one, with AI sitting atop two already-established foundations.

The first wave is data processing—the foundational engine powering modern economies. Banks, e-commerce platforms, and credit card networks process enormous datasets containing customer information, costing hundreds of billions annually and demanding accelerated computing just to function at scale.

The second wave emerged over the past 15 years: recommender systems. Those algorithms personalizing your social media feed, suggesting products, and targeting advertisements have become a dominant computing workload. This demonstrates how AI is already woven into daily economic activity rather than being some speculative future technology.

Now comes the third wave: agentic AI. Autonomous agents like ChatGPT, Grok, and Gemini Space Station Inc. (GEMI) represent a shift from general-purpose computing to specialized accelerated computing, according to Huang. But here's the key difference from previous technology bubbles—the hardware infrastructure supporting this transformation is already deployed and generating real returns.

So when Huang pushes back on the bubble narrative, he's not making a valuation argument. He's pointing to an irreversible shift in how computing works, driven by physical constraints in chip technology and evidenced by trillions in committed capital. Whether that means Nvidia stock is fairly valued is another question entirely, but the underlying technological shift appears to be very real.