Andi Mann, Apica’s CPTO, breaks down research showing that telemetry costs from agentic AI have already outpaced compute costs in 69% of enterprise deployments, with 59% of organizations killing or pausing AI projects because the monitoring bills got too big. The agents getting shelved first are the ones covering cybersecurity, compliance, and fraud detection. Mann’s argument is that a pipeline-first observability architecture, one that routes and filters data before it hits the platform, is the only way to stay ahead of the 9.5x average data volume growth agentic AI brings.
- Article
Agentic AI’s Hidden Cost Crisis: How Telemetry Overload Is Killing Enterprise AI Deployments Before They Scale – VMblog QA
