Corporate America faces AI hangover as spending surges to $2.5T

Global enterprise AI spending is forecast to hit roughly $2.5 trillion in 2026, a 44 to 47% jump over 2025 levels, according to Gartner projections. To put that in perspective, the entire GDP of France sits around $3 trillion.

The single largest driver, according to research circulated by Fortune, is not a calculated return-on-investment thesis. It is fear. Specifically, the fear that a competitor is moving faster.

Everyone is buying. Not everyone is winning.

Organizations are not shuffling money from old budgets. Many are creating entirely new budget lines just for generative AI tools: Microsoft Copilot, Google Gemini, Anthropic’s Claude, and their enterprise variants.

According to PwC’s 2026 CEO Survey, only about 25% of AI initiatives are delivering expected returns. A separate survey found that 56% of business leaders reported zero revenue gains or cost savings from their AI implementations in the prior year.

Big Tech is cashing the checks regardless

Alphabet, Amazon, Meta, and Microsoft collectively carry roughly $2.4 trillion in AI-related commitments, and their projected capital expenditure for AI infrastructure in 2026 alone sits between $700 billion and $800 billion.

AI infrastructure spending is expected to account for approximately $1.37 trillion of total 2026 AI spend. Nvidia has projected that longer-term annual AI spending could reach $3 to $4 trillion by 2030.

The hangover is starting to show

Starting in September 2026, companies including TIAA and Carvana began implementing token limits on employee AI usage, a direct attempt to rein in costs that had grown faster than the value they produced.

What makes this cycle different is the scale. Previous enterprise software waves, ERP rollouts in the nineties, cloud migration in the 2010s, cost tens or hundreds of billions cumulatively. This one is clearing $2.5 trillion in a single calendar year.

Vendors whose products sit in the 75% of AI initiatives not delivering returns are already facing harder renewal conversations. Enterprise software companies that bundled AI features into existing contracts at premium prices will face mounting pressure to demonstrate value before the next renewal cycle.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.