AI Is the World’s Most Powerful Utility and We Need to Treat It as Such

Artificial intelligence is the world’s most powerful utility, and we need to treat it as such.

Like electricity, its economic value will come from what people and businesses accomplish once it is applied across the economy.

That expectation supports some of the largest valuations in the world. Investors are not merely assigning value to models, data centers or chatbots. They are pricing the belief that AI will improve nearly every productive industry.

I saw the clearest expression of that argument while reading the SpaceX prospectus. SpaceX estimates that enterprise applications represent $22.7 trillion of a projected $26.5 trillion AI market. Goldman Sachs projects that the company’s AI revenue could increase one hundredfold by 2030—roughly 10,000 percent in five years.

The most important word in those projections is applications. Nearly all the anticipated value comes not from the existence of artificial intelligence, but from its deployment throughout the economy.

For now, the expectation is enough to support extraordinary valuations. Over time, the applications have to produce the economic value those valuations anticipate.

The early evidence is uneven. AI is already highly effective where the buyer is concentrated, the budget is large and the objective is clear. Its most advanced applications include surveillance, targeting, autonomous weapons.

These uses prove that the technology works. They do not demonstrate the broad economic transformation reflected in the valuations.

Artificial intelligence has not cured a major disease, made healthcare broadly affordable or produced a measurable decline in the cost of living. Productive applications are emerging in software development, customer service, fraud detection and logistics, but they remain early examples of a much larger claim.

The enterprise thesis extends beyond making existing digital work faster. A doctor can detect disease earlier. A manufacturer can reduce waste. An individual can build a company that once required hundreds of employees and millions of dollars.

When these applications spread, artificial intelligence begins to function like electricity.

Electricity did not become the economy. It powered the economy. It moved through factories, hospitals, farms and homes. The electric company sold the power, while the larger economic value appeared in everything people built with it.

Artificial intelligence must operate the same way. The companies building the models can prosper by supplying intelligence, while the larger gains emerge among the people and industries applying it.

AI differs from electricity in one critical respect. Electricity was generated from natural resources. Artificial intelligence was trained on our human knowledge. We are the source of AI.

Its models absorbed books, journalism, scientific research, software, art and historical records. Millions of people supplied the intellectual foundation. A small number of companies now control the systems that organize it.

That structure can produce enormous technology companies. A more productive economy emerges when the intelligence moves outward and increases the capabilities of everyone using it.

The market has priced the promise. But the whole shooting match will be revealed in the deployment.

Electricity became valuable because it made everything else more valuable. Artificial intelligence now faces the same test.