Nvidia CEO Jensen Huang: Why AI is Not a Bubble

Addressing the AI Bubble

• 💬 Nvidia CEO Jensen Huang addresses concerns about a potential AI bubble, urging a return to first principles in computer science.

The End of Moore's Law

• ⚠️ Moore's Law has run its course, leading to challenges in meeting the demand for computing with general-purpose processors.
• 📈 The world is rapidly transitioning to accelerated computing to overcome these limitations, a trend Nvidia has championed for over 20 years.

Rise of Accelerated Computing

• 📊 In just six years, the proportion of CPUs in top 500 supercomputers dropped from 90% to less than 15%, while accelerated computing rose from 10% to over 90%.
• 💰 Hundreds of billions of dollars are spent on raw data processing (e.g., SQL, data frames) which is now shifting to accelerated computing, driving significant demand.

Generative and Agentic AI

• 💡 Generative AI, exemplified by recommender systems (Rexus), is a crucial application that is transitioning from CPUs to GPUs.
• 🚀 Agentic AI (e.g., Grok, OpenAI, Anthropic, Gemini) represents a third layer built upon these foundational shifts in computing and generative AI.

Justifying AI Investment

• ✅ The massive underlying movement from general-purpose to accelerated computing, fueled by data processing and generative AI, provides a strong justification for the investment in revolutionary agentic AI.
• 🎯 This fundamental shift suggests that the resources required for agentic AI are substantially less than perceived and are entirely justified by the technological evolution.