AI Inference Token Capacity Planning

Help enterprise customers plan AI inference capacity by converting an annual Token target into a concrete supply plan, SLA commitments, and cost boundaries.

SCOPE: Analysis and recommendation only. This skill does NOT provision resources or modify infrastructure.

RELATED SKILL: If the deployment mode is on-premise GPU clusters, use container-management:gpu-bottleneck-analysis to verify actual cluster capacity after the plan is drafted.

Interaction Gate

Before any calculation, recommendation, or final structured report, check whether the user has provided enough planning inputs.

MUST ask clarifying questions first when any of these required inputs are missing: