AI Compute Demand Growth and the Infrastructure Supply Constraint

The discussion examines whether current AI investment resembles a speculative bubble by comparing AI capacity building with the 1990s telecom infrastructure boom. It focuses on the mismatch between rapidly rising demand for AI models and the slower delivery of compute, power, chips, and capital needed to support deployment.

- Contrast between the 1990s telecom buildout and AI capacity expansion
- Demand trends cited from portfolio companies, including model spending growth, customer demand for AI deployment, and personalized agents
- Supply-side bottlenecks involving entitlements, power, chips, and capital formation
- Infrastructure investment focus on long-term energy assets and data centers contracted to large companies

For viewers seeking a structured overview of AI infrastructure economics, the discussion provides a framework for assessing demand sustainability, supply constraints, and the role of long-term contracted assets. It can help inform due diligence questions and strategic planning around AI compute capacity.