For the past two years, investors have viewed artificial intelligence primarily through the lens of software and semiconductors.
The focus has been on large language models, breakthrough applications, and the companies designing the chips that make those systems possible.
That narrative has created enormous wealth.
Nvidia became one of the most valuable companies in the world. Microsoft, Alphabet, Amazon, and Meta committed hundreds of billions of dollars to AI infrastructure. Venture capital poured into every corner of the ecosystem.
Yet a more important story is beginning to emerge.
The next phase of the AI boom may have less to do with algorithms and more to do with the physical systems that support them.
Electricity, transmission networks, transformers, industrial sites, and critical minerals are becoming strategic assets in the global competition for AI leadership.
One of the clearest signals arrived this month from an unexpected place.
At President Emmanuel Macron's Choose France summit, SoftBank backed plans to invest roughly €45 billion in three AI data centers in northern France. The project will eventually deliver 3.1 gigawatts of capacity.
What stood out was not the size of the investment. It was the reason behind it.
SoftBank executives pointed directly to France's ability to produce and export electricity, including power generated by its extensive nuclear fleet. Access to reliable energy was a decisive factor in the company's decision.
That should get investors' attention because it reveals how the economics of AI are evolving.
Capital is increasingly flowing toward power.
The Physical Side of AI