Nvidia CES 26 Analyst Q&A, CEO Jensen Huang and CFO Colette Kress

The podcast centers on questions regarding NVIDIA's technology, strategy, and market outlook. NVIDIA's CEO addresses inquiries about Vera Rubin's production, Grok licensing, and autonomous vehicle market scope, emphasizing the company's focus on extreme co-design to overcome Moore's Law. He highlights NVIDIA's unique position in the supply chain, being the only chip company that buys DRAM directly, and its partnerships with every cloud provider. The discussion also covers the company's approach to DGX Cloud, its role in open-source AI, and its strategy for maintaining high margins by delivering value through training, inference, and overall data center efficiency. The podcast further explores the potential of agentic AI and physical AI, and NVIDIA's investments across various layers of the technology stack.

Outlines

Part 1: Benchmarking and System Co-design

Part 2: Market Strategy and Industry Verticals

Part 3: Supply Chain and Infrastructure

Part 4: Technical Challenges and Hardware Evolution

Part 5: Agentic AI and Future Applications

Part 6: Ecosystem and Business Economics

Part 7: R&D, M&A, and Architectural Outlook

Sign in to continue reading, translating and more.

Continue

NoRush Invest

Part 1: Benchmarking and System Co-design

Comparing Rubin's Performance to TPUs and MLPerf Benchmarking Rigor

Semi-Analysis Benchmarks, Extreme Co-design, and the Key to Overcoming Moore's Law

NVIDIA's ASICs, Vera Rubin's Production, and Grok Licensing for Low Latency AI

Part 2: Market Strategy and Industry Verticals

Vera Rubin's Production Timeline, Autonomous Vehicle Market Scope, and Revenue Models

NVIDIA's Approach to Insanely Hard Problems, Anthropic Collaboration, and AI Platform Strategy

Part 3: Supply Chain and Infrastructure

Supply Chain Dynamics, Power Considerations, and NVIDIA's Upstream and Downstream Advantages

China Supply, DGX Cloud Strategy, and Customer Attraction for CSPs

Context Memory Storage Control, Bluefield, and Vera Rubin's Ramping Velocity

Part 4: Technical Challenges and Hardware Evolution

Vera Rubin's Unique Challenges, De-risking Process, and the End of Moore's Law

Ramping Vera Rubin, NVLink Hot-Swapping, and the Frontier Model Market

Part 5: Agentic AI and Future Applications

Vertical Agentic Systems, Foundation Model Companies, and Open Models

Agentic AI, Application Frameworks, and the Future of AI Systems

Alpamayo, Physical AI Market Sizing, and Multimodality

Part 6: Ecosystem and Business Economics

NeoClouds, GPU Clouds, and Regional AI Infrastructure

Margin Sustainability, Value Delivery, and the One Gigawatt Data Center

Game Changers for AI Models, Token Growth, and the Next Milestone for NVIDIA

Part 7: R&D, M&A, and Architectural Outlook

Extreme Code Design, R&D Intensity, and M&A Philosophy

Grok's Software Technology, NVIDIA's AI Infrastructure, and HBM Usage

CPX Benefits, Workload Versatility, and the Future of AI

Nvidia CES 26 Analyst Q&A, CEO Jensen Huang and CFO Colette Kress

NoRush Invest

Part 1: Benchmarking and System Co-design

00:00Comparing Rubin's Performance to TPUs and MLPerf Benchmarking Rigor

Comparing Rubin's Performance to TPUs and MLPerf Benchmarking Rigor

03:37Semi-Analysis Benchmarks, Extreme Co-design, and the Key to Overcoming Moore's Law

Semi-Analysis Benchmarks, Extreme Co-design, and the Key to Overcoming Moore's Law

07:22NVIDIA's ASICs, Vera Rubin's Production, and Grok Licensing for Low Latency AI

NVIDIA's ASICs, Vera Rubin's Production, and Grok Licensing for Low Latency AI

Part 2: Market Strategy and Industry Verticals

12:42Vera Rubin's Production Timeline, Autonomous Vehicle Market Scope, and Revenue Models

Vera Rubin's Production Timeline, Autonomous Vehicle Market Scope, and Revenue Models

17:11NVIDIA's Approach to Insanely Hard Problems, Anthropic Collaboration, and AI Platform Strategy

NVIDIA's Approach to Insanely Hard Problems, Anthropic Collaboration, and AI Platform Strategy

Part 3: Supply Chain and Infrastructure

23:52Supply Chain Dynamics, Power Considerations, and NVIDIA's Upstream and Downstream Advantages

Supply Chain Dynamics, Power Considerations, and NVIDIA's Upstream and Downstream Advantages

29:34China Supply, DGX Cloud Strategy, and Customer Attraction for CSPs

China Supply, DGX Cloud Strategy, and Customer Attraction for CSPs

36:07Context Memory Storage Control, Bluefield, and Vera Rubin's Ramping Velocity

Context Memory Storage Control, Bluefield, and Vera Rubin's Ramping Velocity

Part 4: Technical Challenges and Hardware Evolution

40:30Vera Rubin's Unique Challenges, De-risking Process, and the End of Moore's Law

Vera Rubin's Unique Challenges, De-risking Process, and the End of Moore's Law

44:58Ramping Vera Rubin, NVLink Hot-Swapping, and the Frontier Model Market

Ramping Vera Rubin, NVLink Hot-Swapping, and the Frontier Model Market

Part 5: Agentic AI and Future Applications

47:40Vertical Agentic Systems, Foundation Model Companies, and Open Models

Vertical Agentic Systems, Foundation Model Companies, and Open Models

51:29Agentic AI, Application Frameworks, and the Future of AI Systems

Agentic AI, Application Frameworks, and the Future of AI Systems

53:35Alpamayo, Physical AI Market Sizing, and Multimodality

Alpamayo, Physical AI Market Sizing, and Multimodality

Part 6: Ecosystem and Business Economics

57:34NeoClouds, GPU Clouds, and Regional AI Infrastructure

NeoClouds, GPU Clouds, and Regional AI Infrastructure

1:01:03Margin Sustainability, Value Delivery, and the One Gigawatt Data Center

Margin Sustainability, Value Delivery, and the One Gigawatt Data Center

1:07:45Game Changers for AI Models, Token Growth, and the Next Milestone for NVIDIA

Game Changers for AI Models, Token Growth, and the Next Milestone for NVIDIA

Part 7: R&D, M&A, and Architectural Outlook

1:13:06Extreme Code Design, R&D Intensity, and M&A Philosophy

Extreme Code Design, R&D Intensity, and M&A Philosophy

1:18:06Grok's Software Technology, NVIDIA's AI Infrastructure, and HBM Usage

Grok's Software Technology, NVIDIA's AI Infrastructure, and HBM Usage

1:20:57CPX Benefits, Workload Versatility, and the Future of AI

CPX Benefits, Workload Versatility, and the Future of AI