Within this $64 billion bucket, GenAI models — both foundation and domain-specific — are the high-growth engine, while traditional data science/ML platforms and AI app-development platforms grow at a steadier pace.

The Fastest-Growing Subcategories

Domain-specific language models (DSLMs) and specialized GenAI models are the standout category, with spending more than tripling in a single year. Gartner senior principal research analyst Arunasree Cheparthi noted that organizations are increasingly seeking models tailored to their industry or task, rather than relying solely on general-purpose foundation models .

Foundation GenAI models, while growing at a lower percentage rate (104.2%), still represent a larger absolute increase — from $11.4 billion to $23.4 billion — and remain the single largest spending line within the model category .

How $64 Billion Fits Inside $2.59 Trillion (May 2026)

Gartner issued a broader worldwide AI spending forecast on May 19, 2026, covering the full AI technology stack . The key numbers:

• Total worldwide AI spending (2026): $2.59 trillion, up 47% year-over-year from roughly $1.76 trillion in 2025 .
• AI infrastructure dominates: More than 45% of the total goes to vendor and hyperscaler infrastructure build-out — servers, semiconductors, and data center hardware .
• AI agent software is the fastest-growing overall segment, rising from $86.4 billion in 2025 to $206.5 billion in 2026, about +139% .
• AI software (including platforms, models, and AI-enhanced SaaS) is a subset of this $2.59 trillion; the $64 billion platforms-and-models figure sits inside the broader AI software category, which Gartner estimates at roughly $400 billion+ in total enterprise AI software spending .
• AI services (consulting, implementation, managed services) and AI-embedded devices (GenAI smartphones, etc.) account for the remainder.

The key relationship: The $64 billion "AI platforms and models" market is the end-user spending on core building blocks. It represents roughly 2.5% of the $2.59 trillion total AI spend — because the vast majority of that $2.59 trillion goes to vendor-side infrastructure, AI-enhanced hardware, services, and embedded AI in products (e.g., AI PCs, smartphones) rather than platform/model procurement alone .

Caveat: Two Forecast Updates in 2026

Gartner revised its total-AI spending forecast upward between January and May 2026. The January 2026 forecast pegged total AI spend at $2.52 trillion (44% YoY) ; the May update raised that to $2.59 trillion (47% YoY), adding roughly $70 billion, partly due to higher expected model consumption . The July 2026 platforms-and-models forecast is consistent with that upward revision.

What This Means

The data paints a clear picture: GenAI models, especially domain-specific ones, are the growth engine within AI spending. But the numbers also show that the bulk of AI investment remains in infrastructure — the servers, chips, and data centers that power those models. For enterprises, the takeaway is that model spending is accelerating fast, but infrastructure costs still dominate the overall AI budget.

Within this $64 billion bucket, GenAI models — both foundation and domain-specific — are the high-growth engine, while traditional data science/ML platforms and AI app-development platforms grow at a steadier pace.

The Fastest-Growing Subcategories

Domain-specific language models (DSLMs) and specialized GenAI models are the standout category, with spending more than tripling in a single year. Gartner senior principal research analyst Arunasree Cheparthi noted that organizations are increasingly seeking models tailored to their industry or task, rather than relying solely on general-purpose foundation models .

Foundation GenAI models, while growing at a lower percentage rate (104.2%), still represent a larger absolute increase — from $11.4 billion to $23.4 billion — and remain the single largest spending line within the model category .

How $64 Billion Fits Inside $2.59 Trillion (May 2026)

Gartner issued a broader worldwide AI spending forecast on May 19, 2026, covering the full AI technology stack . The key numbers:

• Total worldwide AI spending (2026): $2.59 trillion, up 47% year-over-year from roughly $1.76 trillion in 2025 .
• AI infrastructure dominates: More than 45% of the total goes to vendor and hyperscaler infrastructure build-out — servers, semiconductors, and data center hardware .
• AI agent software is the fastest-growing overall segment, rising from $86.4 billion in 2025 to $206.5 billion in 2026, about +139% .
• AI software (including platforms, models, and AI-enhanced SaaS) is a subset of this $2.59 trillion; the $64 billion platforms-and-models figure sits inside the broader AI software category, which Gartner estimates at roughly $400 billion+ in total enterprise AI software spending .
• AI services (consulting, implementation, managed services) and AI-embedded devices (GenAI smartphones, etc.) account for the remainder.

The key relationship: The $64 billion "AI platforms and models" market is the end-user spending on core building blocks. It represents roughly 2.5% of the $2.59 trillion total AI spend — because the vast majority of that $2.59 trillion goes to vendor-side infrastructure, AI-enhanced hardware, services, and embedded AI in products (e.g., AI PCs, smartphones) rather than platform/model procurement alone .

Caveat: Two Forecast Updates in 2026

Gartner revised its total-AI spending forecast upward between January and May 2026. The January 2026 forecast pegged total AI spend at $2.52 trillion (44% YoY) ; the May update raised that to $2.59 trillion (47% YoY), adding roughly $70 billion, partly due to higher expected model consumption . The July 2026 platforms-and-models forecast is consistent with that upward revision.

What This Means

The data paints a clear picture: GenAI models, especially domain-specific ones, are the growth engine within AI spending. But the numbers also show that the bulk of AI investment remains in infrastructure — the servers, chips, and data centers that power those models. For enterprises, the takeaway is that model spending is accelerating fast, but infrastructure costs still dominate the overall AI budget.