We think the better question is this: under what conditions does AI become economically real?
To us, AI is not just a software story. It is a productivity story on the business side and an infrastructure story on the energy side.
When AI Becomes Economically Real
AI starts behaving like an actual economic tool — rather than a trend — when it does one or more of four things:
- Reduces labor friction
- Increases response speed
- Improves asset utilization
- Raises output without proportional cost growth
The business case is already large. McKinsey has estimated that generative AI could add US$2.6 trillion to US$4.4 trillion annually across the use cases it analyzed.
AI Without Energy Is Theory
As AI scales, the system needs more electricity, more grids, more firm capacity, more cooling, and better power reliability. That is why AI and energy are increasingly part of the same conversation.
The IEA projects that global electricity demand from data centres could roughly double to around 945 TWh by 2030, with AI as the main driver of that growth.
Fast-deploying renewables will likely be part of the solution, especially where they can add capacity quickly and sustainably. But the real answer is broader than renewables alone: grids, storage, cooling, firm power, and reliability will matter just as much.
Our Takeaway
AI is not economically real because it sounds transformative. It becomes economically real when it improves productivity and when the underlying infrastructure can carry the load. That is where the most important business and investment conversations are heading.
“AI is not economically real because it sounds transformative. It becomes economically real when it improves productivity and the underlying infrastructure can carry the load.”
