NVIDIA (NVDA) CES 2026 Keynote summary

Industry platform shifts and AI transformation

• Two simultaneous platform shifts: applications now built on AI, and software development fundamentally changing from programming to training models.

• Every layer of the computing stack is being reinvented, modernizing trillions in computing investments and shifting global R&D budgets toward AI.

• Open models and open innovation have accelerated AI proliferation, with rapid advances in reasoning systems and agentic models.

• Agentic AI systems now reason, plan, use tools, and interact with the physical world, enabling new applications and industries.

• Open-source models and libraries empower global participation in the AI revolution, with top performance in multimodal and retrieval tasks.

AI model advancements and scaling

• Open models are rapidly reaching the frontier, with significant releases from 2023 to 2026, and 80% of startups are building on open models.

• Model size and parameter count are growing 10x per year, with token costs dropping 10x annually, driving increased adoption and capability.

• AI is evolving beyond large language models, integrating physical reasoning and agentic capabilities, and learning laws of nature.

• Monthly downloads of open models on platforms like Hugging Face are surging, reflecting widespread industry adoption.

• Open models now generate a significant share of tokens, with 1-in-4 OpenRouter tokens produced by these models.

AI applications, agentic frameworks, and enterprise integration

• AI applications are now built as agentic frameworks, combining proprietary and customized models for reasoning and tool use.

• Enterprises integrate agentic AI systems as user interfaces, transforming platforms like Palantir, ServiceNow, and Snowflake.

• Physical AI and AI physics enable AIs to understand and interact with the physical world, using simulation and synthetic data for training.

• Omniverse and Cosmos provide digital twin and world foundation models, supporting robotics and autonomous systems.

• Demonstrations showed customizable AI assistants and robotics, highlighting rapid development enabled by open-source tools.