Nvidia’s Jensen Huang says AI boom is just starting, rejects bubble talk

Nvidia CEO Jensen Huang

In the run-up to Nvidia’s third-quarter earnings, investor chatter circled around a single anxiety: are we in an AI bubble that’s about to burst? With tens of billions pouring into new data centres, many on Wall Street questioned whether the returns could ever match the hype. Nvidia CEO Jensen Huang used his opening remarks on the earnings call to knock that theory down immediately.

Huang said he sees “something very different” from a bubble, and his view carries weight. Nvidia is the spine of the AI industry, powering every major cloud platform from Amazon and Microsoft to Google and Oracle. It also counts the biggest model developers — OpenAI, Anthropic, xAI and Meta — among its most enthusiastic customers. Few executives have a clearer view of global GPU demand.

Huang offered a three-part argument to explain why the current surge is neither temporary nor overstated. First, he said large swathes of existing computing — data processing, ad recommendations, search, engineering workloads — are shifting to GPUs because traditional CPU-based systems simply can’t cope with modern AI requirements. That transition alone represents a massive long-term upgrade cycle.

Second, he argued that AI won’t just optimise today’s apps but will unlock entirely new categories of software built from the ground up around intelligence. These applications, he suggested, will rely on sustained and exponential compute growth.

His third point centred on “agentic AI”. These AI systems will handle tasks end-to-end, reasoning and planning with minimal user input. Huang said this class of software will be even more compute-hungry, requiring significant GPU-led infrastructure.

Nvidia, he stressed, is the only company equipped to address all three fronts. “Each will contribute to infrastructure growth in the coming years,” he said, urging investors to look at the long-term fundamentals rather than bubble rhetoric.