Table of Contents
Gartner’s March 2026 report, 10 Best Practices for Optimizing Generative and Agentic AI Costs (Gartner, 10 Best Practices for Optimizing Generative and Agentic AI Costs, Arun Chandrasekaran et al., 20 March 2026), examines the financial and operational realities of scaling generative and agentic AI in the enterprise.
Airia was included as a sample vendor in the section discussing AI gateways.
As organizations move from experimentation to production deployments, cost management becomes central to long-term AI viability.
Gartner writes: “Through 2028, at least 50% of GenAI projects will overrun their budgeted costs due to poor architectural choices and lack of operational know-how.”
The report also states: “Organizations transitioning from GenAI pilots to production experience a rude awakening when it comes to costs.”
From our perspective, this shift reflects the difference between experimentation and enterprise systems engineering. Production AI introduces sustained inference usage, workflow automation, integration requirements, governance controls, and lifecycle management obligations that do not surface during limited pilots.
Agentic AI and Cost Escalation
Agentic systems introduce additional complexity because they orchestrate chains of actions rather than single responses. In the report, Gartner notes: “Agents trigger chains of actions, not single actions, which means a single user request can balloon into tens or hundreds of LLM calls.”
As AI agents become embedded into workflows, cost dynamics become nonlinear. Recursive processes, tool usage, and multi-step execution can increase token consumption and infrastructure demand quickly if not governed systematically.
Gartner further advises: “Establish strict usage and access governance: Implement quotas, model guardrails, and monthly utilization reviews to prevent consumption sprawl and surprise costs.”
In our view, this reinforces that cost control is not simply a budgeting exercise — it is a governance and architecture issue.
AI Gateways and Enterprise Control Planes
In its discussion of automated model selection, caching, and routing, the report introduces AI gateways: “A new category of tools called AI gateways can help control costs by enforcing policies to track and manage access to AI services and by providing features such as caching and model routing to reduce costs.”
The report adds: “In addition, AI gateways can provide usage-based cost tracking and reporting based on AI tokens, rate limiting and quota enforcement to prevent excessive or unintended usage.”
Airia was listed among vendors offering AI gateway tools in this section.
From our perspective, centralized control over model access, routing logic, and policy enforcement becomes increasingly important as enterprises engage multiple model providers and deploy agents across departments.
Architectural and Lifecycle Considerations
Beyond gateways, the report emphasizes disciplined decision-making around model selection, customization, and hosting strategies. It warns that cost pressures may reshape long-term AI strategies: “By 2028, more than 50% of enterprises that have built their own models from scratch will abandon their efforts due to costs, complexity and technical debt in their deployments.”
The report also states: “Through 2028, the aggregated costs of model inference will be at least 70% of the total model lifetime costs, eclipsing the training costs by a considerable margin.”
These statements underscore that inference, routing, orchestration, and operational governance represent persistent cost centers in production environments.
From our point of view, sustainable AI scale requires structured orchestration, lifecycle oversight, shared infrastructure patterns, and ongoing visibility into usage and cost drivers.
From Pilot Enthusiasm to Operational Discipline
The report makes clear that production AI requires different controls than pilot experimentation: “Creating a production-ready GenAI system can be orders of magnitude more expensive than running a pilot. Every token counts — strive to minimize costs across the AI life cycle.”
In our opinion, organizations that treat AI as a coordinated platform capability — rather than a collection of disconnected experiments — are better positioned to manage these lifecycle costs.
Conclusion
Gartner’s March 2026 report, 10 Best Practices for Optimizing Generative and Agentic AI Costs, outlines operational and financial considerations for enterprises scaling generative and agentic AI.
Airia was included in the section discussing AI gateway tools.
To learn more about Airia’s approach to AI orchestration, governance, and cost management, connect with our team.
Citation: Gartner, 10 Best Practices for Optimizing Generative and Agentic AI Costs, Arun Chandrasekaran et al., 20 March 2026.
Gartner does not endorse any company, vendor, product or service depicted in its publications, and does not advise technology users to select only those vendors with the highest ratings or other designation. Gartner publications consist of the opinions of Gartner’s business and technology insights organization and should not be construed as statements of fact. Gartner disclaims all warranties, expressed or implied, with respect to this publication, including any warranties of merchantability or fitness for a particular purpose.
