• Consider AI compute and power infrastructure as a long-term investment theme: expanding enterprise use of AI agents may increase demand for both, while supply constraints could favor providers with available capacity.
• No specific publicly traded company, ticker, price target, or timeframe is identified, so treat this as a sector theme rather than a buy recommendation.

AI Compute and Power Infrastructure

• The speaker says compute capacity has been bought up by quantitative trading firms and AI companies such as OpenAI and Anthropic, leaving many other companies unable to secure enough capacity.
• The transcript presents power supply as a potential bottleneck for future compute buildout. It gives the example of an executive proposing a $1 billion prepayment for “Chamath’s next power supply” to secure capacity.
• The mention does not identify a specific power project, company, or publicly traded security.

Takeaways

• The discussion points to a possible investment theme in AI infrastructure, particularly compute capacity and the power needed to support it.
• For investors or businesses exploring the theme, the transcript suggests looking at how access to power and compute could affect the ability to meet future demand. The example is an executive funding proposal, not a specific investment recommendation.

Enterprise AI Agents

• The speaker expects companies in insurance, manufacturing, and other industries to need hundreds of thousands or even millions of AI agents to remain competitive.
• That expected demand is presented as a reason companies will need more compute—and therefore more power infrastructure.

Takeaways

• The transcript’s investment case is indirect: demand for enterprise AI agents could increase demand for compute and power capacity.
• The speaker identifies supply constraints as a challenge: compute is described as “sold out,” and companies may struggle to secure the capacity they need.