Nvidia says AI demand is going vertical

Demand, but make it theatrical

Nvidia’s latest earnings call came with the kind of confidence only a company sitting in the middle of the AI gold rush can muster. Jensen Huang said demand has “gone parabolic,” which is CEO-speak for: everyone wants more AI compute, yesterday.

The new AI pitch: compute = cash

Huang’s big thesis is that agentic AI has arrived, meaning models aren’t just generating cute chatbot answers anymore — they’re doing useful work that customers will pay for. In Nvidia land, that means compute isn’t a cost center. It’s the product, the revenue engine, the whole enchilada.

He also argued Nvidia sits at the center of a giant AI rebuild:

• frontier model training
• hyperscale cloud spending
• sovereign AI projects
• enterprise AI infrastructure
• robotics and other physical AI use cases

That’s a lot of buckets, and Nvidia wants a pipe into all of them.

Why the rest of the AI crew matters

The article also name-checks a handful of stocks that could ride the same wave. Think of them as the supporting cast in Nvidia’s blockbuster sequel:

• AMD for alternative accelerators
• Broadcom for custom chips and networking
• Marvell for interconnects and optics
• Micron for memory and HBM
• Super Micro for AI server racks

Big picture: Nvidia is still telling Wall Street the AI buildout is nowhere near done. If Huang’s right, the winners won’t just be chipmakers — they’ll be anyone selling the plumbing that keeps the AI factory humming.