WSJ: AI spending-revenue gap hits a record $1.5 trillion

The gap between what tech companies are spending on artificial intelligence and the revenue generated by AI products has reached $1.5 trillion, according to The Wall Street Journal.

Sequoia Capital venture capitalist David Cahn first flagged the imbalance three years ago, calling it AI’s “$200 billion question.” The gap has widened rapidly since then, reaching $600 billion in 2024, $840 billion in 2025 and more than $1.5 trillion in 2026.

Even Nvidia’s record-breaking growth is not enough to close it. Analysts expect the chipmaker to generate about $400 billion in revenue this fiscal year, rising to $570 billion next year and $700 billion by 2028. But building data centers typically costs roughly twice as much as the chips themselves, before software and cloud provider markups are factored in. Cahn estimates that supporting Nvidia’s projected revenue two years from now would require around $2.1 trillion in end-market demand for AI products.

OpenAI is already generating $27 billion in annual revenue, while Anthropic has passed $46 billion. Google, Amazon and Microsoft are also reporting growing AI-related revenue, but the combined figures still fall well short of the spending required to build out the infrastructure.

The current boom is effectively betting that demand will eventually catch up with investment. The bigger risk, according to experts cited by the Journal, is that investors could run through their available cash and credit before that happens.