Amazon and Google win the AI ​​capex race, but what’s the prize?

Sometimes, it can feel like the AI ​​industry is racing to see who can spend the most money on data centers. He believes that whoever builds the most data centers will have the most computing power, and thus be able to build the best AI products, which will ensure victory in the coming years. There are limits to this way of thinking – traditionally, companies ultimately succeed through manufacturing more Money and spending less — but it has proven remarkably compelling to big tech companies.

If this is the game, Amazon appears to be winning.

The company announced in its earnings on Thursday that it expects $200 billion in capital expenditures throughout 2026, across “artificial intelligence, chips, robotics and low-Earth orbit satellites.” This is higher than the $131.8 billion in capital expenditures in 2025. It is tempting to attribute the entire capital expenditure budget to AI. But unlike most of its competitors, Amazon has a large physical factory, some of which is being converted for use by expensive robots, so non-AI expenses aren’t easy to overlook.

Google is close behind. In its earnings on Wednesday, the company forecast capital expenditures of between $175 billion and $185 billion for 2026, up from $91.4 billion the previous year. It’s much more than the company spent on fixed assets last year, and much more than most of its competitors.

Meta, reported last week, forecasts capital spending of $115 billion to $135 billion for 2026, while Oracle (once the leading model for AI infrastructure) expects $50 billion. Microsoft doesn’t have an official forecast for 2026 yet, but the most recent quarterly figure was $37.5 billion, which works out to roughly $150 billion, assuming it continues to do so. It’s a notable increase, and one that has investors putting pressure on CEO Satya Nadella — but it still puts the company in third place.

From within the world of technology, the logic here is simple. The revolutionary potential of AI will turn edge computing into a scarce resource in the future, and only companies that control their own supply will be able to survive. But while Google, Amazon, Microsoft, Meta, Oracle and others frantically prepare for the computing desert of the future, their investors are not convinced. Each company saw its stock prices decline as investors refused to commit hundreds of billions of dollars, and companies with higher spending tended to decline more.

Importantly, this is not a problem for companies like Meta that have yet to figure out their AI product strategy. It’s everyone — even companies like Microsoft and Amazon with strong cloud businesses and a straightforward approach to how to make money in the age of AI. The numbers are simply too high to comfort investors.

Investor sentiment isn’t everything – and in this case, it may not do much to change the industry’s mind. If you think AI is about to change everything (and the argument is pretty compelling at this point), you’d be a fool to change course just because Wall Street is getting volatile. But from now on, big tech companies will be under great pressure to underestimate the true cost of their AI ambitions.