My notes from the Jensen Huang keynote

The NVIDIA Jensen Huang keynote was both interesting and entertaining. I heard that they call the NVIDIA GTC the Woodstock of Silicon Valley!!!

Here are my notes, in chronological order. Since I’m not a hardware expert, I’ll skip the amazing hardware that was presented and focus on some key ideas—all of which need further research.

General NVIDIA Strategy

NVIDIA’s approach: partner with everyone possible, especially hardware chip companies here in Silicon Valley. Very cool!

They also announced significant investments to further develop San Jose!

The Future of AI

The future of AI lies in the combination of hardware and software. The software will be free! (They showcased a lot of great open-source tools.)

NVIDIA has developed a suite of free AI libraries specialized for various domains (e.g., CUDA-X).

They provide AI infrastructure for enterprises, the cloud, and robotics.

Stages of AI Development

1. Perception AI →
2. Generative AI →
3. Agentic AI →
4. Physical AI (Robots)

For robots, anything that moves will eventually become autonomous!

The Three Challenges of AI Development

1. How and where do we get the data for AI?
2. How do we train AI models effectively?
3. How do we scale AI systems?

A Paradigm Shift in Storage

The storage industry is undergoing a major transformation: moving from data retrieval to semantic Q&A.

The New AI Computer

The DGX—NVIDIA’s new AI computing platform—has arrived and will soon be integrated across all major brands.

The picture below captures the end of the keynote, where Jensen Huang stands alongside a robot developed through a new partnership between NVIDIA, Disney Research, and DeepMind.