AI Platform Spending Is Growing 63% in 2026. Most of It Is Aimed at the Wrong Bucket.

AI platform spending is set to grow 63.4% this year, according to Gartner — and where that growth concentrates says more about enterprise priorities than the headline rate does. Worldwide spending on AI models and platforms hits $64 billion in 2026, up from $39 billion in 2025, with generative AI model spend growing 117% against a more modest 36.9% rise in platform spend. The fastest-growing line item by percentage is domain-specific language models, jumping from $1.6 billion to $4.9 billion (+210%), while the largest category in absolute dollars remains AI platforms for data science and machine learning, up from $19.4 billion to $26.4 billion. Foundation model spending itself more than doubles, from $11.4 billion to $23.4 billion.

Gartner’s own read on where that money is going comes with a warning attached. Arunasree Cheparthi, the firm’s Senior Principal Research Analyst, says enterprise AI budgets are under greater scrutiny, with spend shifting toward providers that can demonstrate clear value across cost, latency, performance, and reliability — a description that sounds a lot like efficiency spend, not growth spend.

That distinction shows up directly in a separate Gartner survey of 204 CFOs, fielded the same month: 45% say their AI investment leans toward productivity and efficiency, while only 20% say it leans toward improving decision quality. The gap matters — functions that invested in what Gartner calls “Upend” initiatives, projects creating new value propositions, products, or markets, were more than twice as likely to report high realized value from AI than functions that stuck to productivity-only plays.

Put the two findings together and the market-sizing number gets more interesting than the growth rate. A 63% spending increase concentrated in efficiency-oriented platform categories, chased by CFOs who by their own admission are underweighting the investment type Gartner’s data says pays off, is a budget-allocation problem hiding inside a growth story. For anyone advising on where AI spend should go next year, the forecast reads less like “the market is growing” and more like “most of the growth is aimed at the wrong bucket.”