The Week in AI: Mary Meeker Drops 340 Slides on the Revolution

AI Adoption and Infrastructure - User adoption of AI is experiencing exponential growth, similar to early internet adoption [1] - The infrastructure investment in AI, including capex from hyperscalers like Tesla, Meta, and OpenAI, is unprecedented, estimated at a $225 billion annual run rate [1] - Token costs (the cost to create a piece of data by AI) have collapsed, reflecting the deflationary nature of the technology [1] Enterprise AI and Open Source - Enterprise AI revenue growth is defying traditional SaaS gravity, with examples like Cursor growing from $1 million to $300 million in ARR in 25 months [1] - Open source AI models are exploding, with significant growth in platforms like Metlam and Hugging Face [1] Physical AI and Automation - Physical AI, particularly robotics, holds immense potential, with synthetic data being used to train robots, potentially including hundreds of millions of humanoid robots [1] - AI-powered systems are being used for large-scale automation, such as weeding 230,000 acres with drones and AI-powered lasers [1] Job Market and Monetization - AI job postings are up 448%, while traditional IT jobs are falling, indicating a significant shift in the job market [1] - Monetization models for AI are still broken, with companies like OpenAI having a $13 billion annual burn rate [1] AI's Transformative Potential - The next 26 billion internet users will be AI-first [1] - The AI revolution is compared to the impact of James Clerk Maxwell's discoveries in electromagnetism, suggesting a seismic transformation for science, technology, and economics [2]