Agentic AI telemetry could raise enterprise monitoring costs

Agentic AI workloads are driving sharp growth in enterprise telemetry, while many organizations say they are unprepared for the resulting data volumes and monitoring costs, according to research by Omdia.

The survey covered more than 300 enterprise IT decision-makers in North America and Western Europe. More than half said their telemetry volume had more than tripled in the previous 12 months, with an average increase of 3.7x. AI and machine-learning workloads accounted for about 43% of that growth.

Looking ahead, respondents expect telemetry from agentic AI workloads to increase by an average of 9.5x within two years. Nearly half anticipate growth of between 6x and 100x. About two-thirds said their organizations were only somewhat prepared or less prepared for the data volumes, while 22% had not considered the implications. At the same time, 35% reported widespread agentic AI deployment.

Among organizations with active agentic AI projects, 69% said observability costs now exceed compute and infrastructure costs. Average annual observability spending was $3.17 million, up 28% year over year. Fifty-nine percent said they had terminated or delayed at least one agentic AI deployment because monitoring costs were too high.

Omdia found that organizations with a telemetry pipeline were 50% more likely to be prepared for data growth of 6x or more. Those with mature pipeline deployments were 80% more likely to have avoided operational cost challenges. The source article reports that 97% have implemented or are considering a pipeline solution, and 70% are actively evaluating one.