Spending on AI models is set to hit $64 billion this year, up 63%
A new Gartner forecast puts 2026 spending on AI models and platforms at $64 billion, a 63% jump. The fastest growth is in the models themselves, but budgets are getting stricter.
The research firm Gartner published a forecast this week putting worldwide spending on AI models and platforms at $64 billion in 2026, up 63% from $39 billion in 2025. It is a big number, but the breakdown underneath it tells the more useful story about where the money is actually flowing.
The fastest-growing slice is generative AI models themselves, the systems that produce text, images, and code. Spending on foundation generative models is set to more than double, from $11.4 billion to $23.4 billion. A “foundation model,” roughly, is a large general-purpose model trained on broad data that other tools are then built on top of. An even faster-growing category is specialized models tuned for narrow domains, projected to jump 210%, from $1.6 billion to $4.9 billion. Steadier growth sits in the platforms companies use to build and run AI: data-science and machine-learning platforms rise from $19.4 billion to $26.4 billion.
Here is what is really going on behind the headline. The growth is huge, but the mood has changed. Gartner points out that enterprise AI budgets are now under real scrutiny, with more attention on usage efficiency, cost control, and measurable outcomes. In plain terms: the era of spending on AI just to be seen doing it is fading, and buyers increasingly want proof that a tool earns its keep on cost, speed, and reliability. That is why the same week can bring both eye-watering growth figures and a wave of “cheaper, more efficient” model launches. The two trends are the same story from different angles.
A caveat worth keeping in mind: forecasts are educated guesses, not facts, and analyst firms have both hits and misses. Treat the $64 billion as a direction of travel rather than a promise.
What this means for you: If your job touches technology budgets, the signal is clear. The question in the room is shifting from “should we use AI” to “is this specific AI worth what it costs,” so being able to point at concrete results matters more than ever. If you are just curious, this explains a pattern you have probably noticed: model makers keep competing on price and efficiency, not only on raw power. When the customers get cost-conscious, the products follow. For most of us, that pressure is quietly good news, cheaper and more reliable tools tend to be the result.
Sources
Google's Gemini Spark agent reaches AI Pro users, and it can juggle 15 tasks at once
Gemini Spark, Google's personal AI agent, rolled out to US AI Pro subscribers on July 24. It runs jobs across Gmail, Docs, and other Workspace apps, up to fifteen at a time.
