Gartner sees AI model and platform spending reaching $64 billion

Global end-user spending on AI models and platforms will reach $64 billion this year, up 63% from last year, Gartner said Monday. The forecast points to a larger enterprise budget fight for AI vendors: CIOs are being pushed to fund broader AI use while proving the cost, performance and reliability case for each deployment.

Gartner said technology leaders are putting more scrutiny on AI budgets as usage expands, particularly around efficiency, cost controls and measurable outcomes. The firm expects enterprise buyers to favor providers that can show value across cost, latency, performance and reliability, rather than selling model access as a standalone capability.

Arunasree Cheparthi, senior principal research analyst at Gartner, told CIO Dive that agentic AI is pushing more model spending into platforms. In her view, model vendors are increasingly supplying intelligence to broader workflow platforms, shifting competition toward vendors that can capture more of an enterprise process from start to finish.

Platform vendors gain leverage as model spend gets embedded

The spending shift is relevant for enterprise software companies because AI budgets are less likely to sit in isolated experimentation lines as deployments mature. Gartner’s view suggests more of the money will be tied to platforms that determine when, where and how models are used across business workflows.

That is a better position for software vendors with existing enterprise distribution than for model providers that rely on direct consumption alone. The report did not disclose vendor-level market share, revenue forecasts by provider or how much of the $64 billion will go to model makers versus application and platform companies.

Cheparthi said AI spending is expected to keep rising year over year, according to CIO Dive, but companies have several ways to keep usage from turning into open-ended consumption. Those controls start with contracts, architecture and governance, according to Gartner.

Contracts are becoming a cost-control layer

Gartner said CIOs can use contract terms to reduce budget volatility. Options include fixing token pricing based on input-to-output ratios, setting hard usage limits inside workflows and requiring approval before teams exceed those limits.

Cheparthi told CIO Dive that enterprises should move toward outcome-based or value-based pricing where possible, because consumption-based pricing can create cost uncertainty and may not match actual business value. She also said companies can negotiate for unused tokens to roll over within a defined period, reducing waste from prepaid capacity that expires.

Those provisions matter because many enterprise AI bills scale with activity, not headcount or seat count. Agentic systems can increase consumption by calling models repeatedly inside a workflow, which makes token visibility a finance issue as much as an engineering metric.

Architecture choices can reduce waste

Gartner said AI architecture is another lever for cost control. Enterprises can use multivendor routing to match tasks with the least expensive model that meets the requirement, reserving premium models for work that needs higher capability while using lower-cost or open-source large language models for routine requests.

Cheparthi said companies that measure cost per token and set targets for reducing repetitive prompts or excess consumption are likely to see stronger returns. She added that token usage visibility should become a standard operating metric across workflows.

Gartner also pointed to segmented budgets by department or application, with hard financial limits and automated throttling or escalation once usage thresholds are reached. That approach gives CIOs a way to set boundaries before adoption spreads across teams.

Governance is part of the AI bill

Gartner said governance is often missed as a cost-saving tool. Vendors may include AI premiums inside broader bundles, increasing renewal costs, so companies should involve contract review teams with AI expertise in governance processes.

Cheparthi said that approach can reduce capital leakage, force tighter measurement of outcomes and connect departmental performance goals to AI initiatives that show measurable value. The report did not specify savings benchmarks from these practices, leaving buyers to test whether governance changes can offset the broader rise in AI spending.

This story draws on original reporting from CIO Dive.