Put power and materials on the AI page
Energy, materials, and resourcesTechnology, software, and semiconductorsThe constraint is the strategyJudgment is the scarce asset
The same six weeks that talk about agentic scale also talk about a power crunch, a transformer supercycle, a materials outlook, and cybersecurity that has to answer at AI speed. Physical AI is framed as the value pool leaders are missing.
What the industry is already doing
- Demand for transformers is described as outrunning supply.
- Fuel retail, LNG procurement, and Hormuz aftershocks say the energy system is already being managed as disruption, not as a steady input.
- Inference scale and soaring AI cost mean the model’s economics and the electron’s economics are now one conversation.
Mindset expected
The constraint is the strategy. A roadmap that assumes the grid catches up is the obsolete one.
What to push this quarter
- 1On the next AI business case, add a line for power, equipment lead time, and the control that fails if agents move faster than security.
- 2Name what fails first.
- 3Delay a capability demo if the constraint is the binding one.
Influence
The AI sponsor, the energy or facilities owner, procurement, and the CISO.
“The model is not the scarce thing. The electron, the transformer, or the control is. Which one fails first for us?”
The decision to ask for
Rewrite one AI case so the physical or control constraint is funded in the same decision as the model.
How it stalls
A capability announcement with a footnote that infrastructure will be addressed in a later phase.