1. Ime, Power generation isn't a major blocker and my peers in the DC space are seeing a decreasing interest in funding new capex expansion because there is a bit of a worry of a Telco Bubble 2.0 arising.

2. Robotics should not be bundled with LLMs, but it is absolutely an AI/ML subfield - and in fact, the kind of subfield that is the best example of AI/ML having tangible real world impact. Autonomous UAS was a sci-fi concept 20 years ago, but is an engineering problem that has become a reality today.