According to analyses from Reuters and Stanford University’s AI Index Report from 2025, global investment in AI has already surpassed the inflation-adjusted cost of the Manhattan Project, the Apollo moon landings, and even the International Space Station combined.
And the scary thing? It’s still accelerating.
More compute
It feels like Sam Altman, CEO of OpenAI, is constantly talking about the need for more “compute” to keep pushing the world towards the next generation of AI, which is Advanced General Intelligence (AGI). What he really means is more data centers, which are racks and racks of servers processing AI requests. But data centers are incredibly expensive to build and run, especially in their power demands and cooling requirements. All this means that AI is rapidly becoming one of the largest physical infrastructure buildouts in modern history.
Gigantic server farms packed with specialized chips, enormous cooling systems and huge electricity supplies are being built worldwide right now.
Reuters’ report claims that investors have already poured nearly $1.6 trillion into AI since 2013. That dwarfs the roughly $36 billion inflation-adjusted cost of the Manhattan Project and even the estimated $250–300 billion cost of the Apollo program. And yes, those are today’s costs, adjusted for inflation.
As we know, the Apollo project put humans on the moon for the first time, and The Manhattan Project changed warfare forever. The International Space Station took decades of international cooperation. Yet AI spending has overtaken them all in just a decade, with most of the money coming from private investors.
$650 billion in 2026
Of course, all the big tech companies are involved. Microsoft, Google, Amazon, Meta, et al, are still spending aggressively to build data centers. Bridgewater Associates estimates that Alphabet, Amazon, Meta, and Microsoft alone could collectively spend around $650 billion on AI infrastructure this year.
Some analysts are even beginning to question whether the industry can physically sustain the pace of expansion.
“There is not enough people and resources to build all this,” ABB CEO Morten Wierod told Reuters while discussing the scale of AI investment.
Others are wondering whether the industry is drifting toward bubble territory. According to the Financial Times, banks are scrambling to offload the risks associated with a huge wave of data center debt as the AI infrastructure boom pushes major lenders toward their financing limits.
Something enormous
That doesn’t necessarily mean the AI boom will collapse. But the scale of today’s AI expansion is a stark reminder that the AI race, which is still being played out in courtrooms right now as well as in data centers, is becoming an industrial transformation measured not in apps, but in power grids, factories, chips, land, water, and trillions of dollars.
While humanity still hasn't managed to returned to the moon, and ambition for a manned flight to Mars seems to be waning, for better or worse, we appear to be building something enormous.
Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds.
