Spending on artificial intelligence is now outpacing the buildouts of the railways and the early internet, even after adjusting for inflation. Data centers alone could absorb more than $30 trillion by mid-century, according to industry forecasts. The scale of this commitment is staggering, and it implies a future in which AI reshapes the global economy.
But the cash to pay for all this hardware and power has not shown up. One leading consultancy estimates the AI industry will need about $6 trillion in annual revenue by 2031 to cover the cost of the computing capacity being installed. Existing consumer and business services might generate only a fraction of that, leaving a gap measured in the trillions.
Filling the shortfall would require AI to move far beyond chatbots and image generators. Models would have to displace search engines and sell ads at scale, power autonomous vehicles and drones, run factories and warehouses, and open entirely new markets in drug discovery and energy. Even then, the payoff hinges on productivity gains that many economists say are still hard to see in the data.
Skeptics point out that justifying today’s valuations would demand years of unusually strong productivity growth, well above the long-run trend. Optimists counter that transformative technologies often take decades to show up in the statistics, and that the infrastructure being built now could support gains that arrive later. Until those gains materialize, the gap between spending and revenue will remain the central tension of the AI boom.
#AI #ArtificialIntelligence #TechSpending #Productivity #DataCenters #TechBubble #Innovation #Economy #FutureOfWork #Investing