NVIDIA
A recap of Nvidia’s “Super Bowl of AI”
Nvidia’s GTC is the hottest ticket in town! (If your town is Silicon Valley, that is.)
Last week, over 30,000 people made their way to San Jose’s SAP Center for The Super Bowl of AI, aka Nvidia’s annual conference.
Jensen Huang hit the stage in his trademark leather jacket to unveil Nvidia’s latest tech debuts, all while giving press quotes and posing for selfies. To say he’s moved from CEO to certified celebrity would be an understatement.
And the hype has only grown since Nvidia became a household name. When GTC started in 2009, the conference was organized for a handful of folks to nerd out about GPUs. Now, it’s where the future of AI gets mapped out in public.
But this year, there was a vibe shift. Huang made it clear that agentic AI is going mainstream, and Nvidia plans to power the whole stack.
Major announcements:
Huang debuted the Groq 3 LPU, a specialized single-core chip that turbocharges existing GPUs
Their new rack-scale system, the Vera Rubin, is designed to speed up and streamline agentic AI models
NemoClaw is Nvidia’s enterprise solution for deploying AI agents at scale
The disconnect
Huang predicted a $1 trillion demand for AI through 2027, since the usual suspects — Microsoft, Google, Amazon, and Meta — have committed to building workflows on Nvidia’s stack.
For businesses, this means the ability to integrate agents into corporate infrastructure, no holds barred. For the rest of us, it points toward more real-time AI support in everyday life.
Sounds promising, right?
You’d think! But Nvidia stock price dropped last week during the conference. So either Wall Street is playing hard-to-get, or the market is less than impressed with GTC’s rollout.
Either way, the math isn’t mathing.
Ah, yes. Silicon Valley’s favorite spectator sport.- TL