Developed market technology companies are spending more than ever on capital projects relative to their revenue. The CapEx-to-sales ratio for these firms hit a record 12% in the latest quarter, a figure analysts say marks the start of a new phase in corporate investment centered on artificial intelligence.
A spending milestone
The ratio measures how much money companies put into long-term assets—data centers, servers, networking gear—for every dollar of sales they bring in. At 12%, the figure has never been higher. The previous peak came during the dot-com buildout of the late 1990s, but today’s surge is being driven by a single technology: AI.
Tech giants and smaller players alike are racing to build the infrastructure needed to train and run large language models and other AI systems. That means buying billions of dollars worth of specialized chips from suppliers like Nvidia, constructing new cloud data centers, and laying fiber optic cable to handle the traffic. The spending is showing up in quarterly earnings reports as capital expenditure lines swell even as revenue growth remains modest.
AI's capital appetite
The jump to a double-digit ratio reflects a strategic bet. Companies are front-loading investment in AI capacity, hoping to capture future revenue from AI-powered products and services. The spending is not limited to the usual suspects—cloud providers and software firms—but also includes companies in sectors such as automotive, healthcare, and finance that are deploying AI internally.
Some investors have begun to question whether the spending will pay off. The ratio of 12% means that for every $100 in sales, $12 goes to capital expenditure. That leaves less cash for dividends, buybacks, or research into other areas. But so far, the consensus among executives is that the risk of underinvesting in AI outweighs the risk of overspending.
What comes next
The record ratio is likely to climb further. Several large tech firms have announced plans to increase their capital budgets for the next fiscal year. The race to build AI infrastructure shows no signs of slowing as companies compete for market share in what they see as the next computing platform.
Earnings calls in the coming weeks will provide more color on how long companies expect to sustain this level of spending. If revenue growth picks up, the ratio could stabilize or even decline. If it doesn't, the 12% figure may become a liability. Either way, the record is a clear signal: the AI era is being built on a scale of capital deployment not seen in decades.
