Where the Value Moves: How Do You Win When Software Is No Longer the Moat?
Mar 3, 2026
Alex Watkins, Citi’s tech capital markets lead who finances data centers and GPUs, and Tony Kim, BlackRock’s technology chief investing across AI infrastructure and models, discuss where value is shifting. They explore the three-layer AI stack, the rise of infrastructure and cost as moats, agentification and robotics, and bold bets like space-based data centers, quantum and timing signals to watch.
AI Snips
Chapters
Transcript
Episode notes
Using AI To Connect Cross‑Domain Research
- Tony Kim uses AI to aggregate disparate ideas across topics, like linking energy, compute and timelines for research.
- He doubts current AI can produce breakthrough novelty like special relativity or an original 1970s screenplay yet.
Value Shifting To Compute Infrastructure
- AI's value is shifting down the stack from software to physical infrastructure because intelligence scales with compute and power.
- Tony Kim frames a three-layer stack: energy/power, chips/data centers, then models/data, and value is moving to that base layer.
Big Gap Between Frontier AI And Everyday Use
- There's a persistent gap between state-of-the-art AI capabilities and what the average person or enterprise actually uses, creating multi-year diffusion dynamics.
- Tony Kim notes state-of-the-art becomes average over time, driving adoption lags and ongoing infrastructure build-out.
