Summary of NVIDIA CEO Jensen Huang Keynote at CES 2025

The keynote by NVIDIA CEO Jensen Huang at CES 2025 introduced groundbreaking advancements in artificial intelligence (AI) and graphics processing units (GPUs). The presentation emphasized the transformative role of "tokens" in AI, which serve as the building blocks for various applications, including image and video generation, robotics, and data analysis.

Key Technological Concepts:

Tokens in AI: Tokens are fundamental units that allow AI to process and generate information across different modalities (text, images, sounds).
Generative AI: The evolution from perception AI to generative AI, which creates new content and understands complex data, was highlighted. This includes advancements in agentic AI, which can perceive, reason, plan, and act.
Transformers: The introduction of Google's Transformer model revolutionized AI, enabling significant advancements in machine learning and computing.
Blackwell Architecture: NVIDIA's latest GPU architecture, the RTX Blackwell family, features:
- 92 billion transistors and enhanced performance capabilities (4,000 TOPS, 4 petaflops of AI).
- New techniques such as neurot texture compression and neuromaterial shading for improved graphics rendering.
- AI-driven technologies like DLSS (Deep Learning Super Sampling) that enhance real-time graphics performance.

Product Features:

RTX Blackwell GPUs: These GPUs are designed to handle extensive AI workloads and graphics processing, with models starting from the RTX 5070 up to the RTX 5090, offering significant performance improvements at competitive price points.
AI Computing Platforms: NVIDIA introduced its AI platforms, including NVIDIA Nims and Nemo, which facilitate the development and deployment of AI agents across various industries.
Cosmos: A new world foundation model designed to understand physical dynamics, enabling advancements in robotics and simulation.

Applications and Use Cases:

Industrial Automation: Partnerships with companies like Keon and Accenture aim to integrate AI in warehouse logistics, optimizing robotic fleets through digital twins and simulations.
Autonomous Vehicles: NVIDIA's next-gen AV computer, Thor, boasts 20 times the processing power of its predecessor, designed for advanced self-driving capabilities.
General Robotics: The introduction of NVIDIA Isaac Groot, a platform for developing humanoid robots, which utilizes synthetic data generation to enhance learning from limited real-world demonstrations.

Future Vision:

Huang articulated a vision where AI becomes integral to everyday computing, with plans to enhance Windows PCs to support AI applications seamlessly. The aim is to democratize AI access and capabilities across various platforms and devices.

Main Speakers:

Jensen Huang, CEO of NVIDIA

Overall, the keynote showcased NVIDIA's commitment to pushing the boundaries of AI and GPU technology, with a clear focus on real-world applications and transformative potential across multiple industries.

Notable Quotes

— 03:10 — « AI is not just a new application with a new business opportunity, but it fundamentally changes how computing works. »

— 03:20 — « The future of computer graphics is neural rendering, the fusion of artificial intelligence and computer graphics. »

— 03:30 — « The age of agentic AI is here for every organization. »

— 03:40 — « The critical capability is how to train these robots. »

— 04:10 — « This is Nvidia's latest AI supercomputer, and it's finally called Project Digits. »