Full Transcript
Loading transcript…
Class #2 | MS&E435: Economics of the AI Supercycle Stanford University Spring '26 Apoorv Agrawal
Watch with subtitles, summary & AI chat
Add the free Subkun extension — works directly on YouTube.
- Watch
- Subtitles
- Summary
- Ask AI
Students and investors interested in the economics of AI, venture capital, and the future of technology.
TL;DR
This lecture from Stanford's MS&E435 class discusses the economic implications of the AI supercycle, contrasting it with the software era. It features insights from investors Brad Gerstner and Sunny Madra on AI's impact on GDP, investment, and the fundamental role of compute.
Key Takeaways
- Unlike software, AI applications require significant compute resources, making their distribution costs non-negligible and impacting business models.
- Technological innovation, particularly in areas like AI, is a societal good that accelerates GDP growth, reduces poverty, and improves quality of life.
- Technology companies have historically outperformed non-tech companies in earnings per share growth, a trend expected to be further amplified by AI.
- The AI supercycle is predicted to have an impact 10 times greater than the industrial revolution, but unfolding at 10 times the speed, potentially within a decade.
- Compute power is the fundamental enabler of AI, acting as the 'atomic unit' for processing and generating intelligence.
- Individuals need to become 'bionic' by leveraging AI to deliver exceptional value in the job market, regardless of traditional educational backgrounds.
- The economic landscape is shifting, with AI poised to revolutionize knowledge work, creating immense opportunities and necessitating societal adaptation.
In This Video
- 00:00AI Supercycle vs. Software
The class discusses the AI supercycle, contrasting it with software's near-zero distribution cost. AI requires significant compute.
- 00:33Introducing Brad Gerstner
Brad Gerstner, founder of Altimeter, is introduced. He manages $15 billion and has invested across market cycles.
- 01:15Altimeter's Investment History
Altimeter invested early in Google, mobile, cloud, and software, including Snowflake and GitLab. Now investing in AI leaders.
- 01:44Invest America Initiative
Brad Gerstner started Invest America, a federal legislation to create birth investment accounts for children.
- 03:22Introducing Sunny Madra
Sunny Madra, a serial entrepreneur, has had multiple successful acquisitions by major tech companies.
- 03:55Sunny Madra's Acquisition Trajectory
Madra co-founded companies acquired by Pivotal, Ford, Grok, and Nvidia, leading innovation labs and platforms.
- 05:10Global GDP and Innovation
Global GDP per capita was stagnant for 1800 years, then accelerated. Innovation drives societal progress and quality of life.
- 06:40Technology's Economic Impact
Technology's share of global GDP is growing. Tech companies compound earnings faster than non-tech, a trend AI will accelerate.
- 08:11Compute: The Root of AI
Compute is fundamental to AI. Sunny Madra will discuss Grok and observations on inference in 2023-2024.
- 08:50Grok's Founding and Vision
Grok was founded by Jonathan Ross, creator of Google's TPU. He transitioned from ad optimization to AI hardware.
Questions & Answers
What is the main difference between software distribution and AI distribution?
Software has near-zero incremental distribution cost, while AI applications require significant compute power for each additional user, making distribution costly.
Who is Brad Gerstner and what is Altimeter?
Brad Gerstner is the founder and CEO of Altimeter, an investment firm managing over $15 billion, known for investing across market cycles and in tech companies like Snowflake and OpenAI.
What is Invest America?
Invest America is a federal legislation creating an investment account at birth for every American child, aiming to foster economic independence and ownership.
What is Sunny Madra's background and recent acquisition?
Sunny Madra is a serial entrepreneur whose companies, including Definitive Intelligence, were acquired by Grok and then Nvidia for $20 billion, the latter being Nvidia's largest acquisition ever.
How has global GDP per capita changed over time?
For the first 1800 years, global GDP per capita saw little change, but it began accelerating in the 1800s and 1900s, now doubling approximately every 25 years.
Why are technology companies better investments than non-tech companies?
Technology companies tend to be better investments because they compound earnings per share faster, with the NASDAQ compounding at 15% over the last 10 years compared to 6% for non-tech companies.
Key Terms
- Compute — The processing power required for AI operations, representing the fundamental resource for AI development and deployment.
- Inference — The process of using a trained AI model to make predictions or decisions, which is a key component of AI application costs.
- TPU (Tensor Processing Unit) — Google's custom-designed hardware accelerator for machine learning, crucial for training and running AI models efficiently.