These days, whenever AI comes up—NVIDIA, xAI, all that world—people casually throw around figures in the tens of billions of dollars. It sounds almost surreal.

Then again, maybe this is exactly how early 19th-century Britain must have felt when iron, railways, and chemicals suddenly replaced cotton as the engines of growth. Compared with spinning mills, the money required must have seemed almost from another universe.

A cotton mill could still be financed by the deposits that local gentry left with provincial banks. But blast furnaces were another matter. So were railways running hundreds of kilometers, along with the endless locomotives, freight cars, and passenger coaches they required. At that point, the old financial world was no longer enough. Britain had to open the door wider to joint-stock companies and let capital gather on an entirely new scale.

That is why today’s AI spending spree feels strangely familiar.

.

Still, one question keeps nagging at me.

Why should AI—which is, after all, just digital signals flashing back and forth—consume so much money?

The usual answer is electricity. Vast server farms, giant data centers, and the power plants needed to keep them humming day and night. Fair enough. But will this power consumption really continue forever in a straight upward line? Technology rarely moves that way.

.

What happens if there is a sudden breakthrough—say, a dramatic reduction of electricity consumption in power semiconductors, perhaps using synthetic diamond or some other material we are barely talking about today? What if the power needed for AI suddenly falls by one order of magnitude? Or two? The entire landscape would change overnight.

.

After all, the human brain consumes only a tiny amount of power. If nature can produce intelligence so economically, it is hard to believe that artificial intelligence must always remain this hungry for electricity. One suspects that sooner or later the real revolution will come not from making AI “smarter,” but from making it vastly more efficient.

And when that day comes, today’s eye-popping investment figures may begin to look less like madness and more like the cost of building the first bridges into a new industrial age.